‘Seeing’ chemistry: investigating the contribution of mental imagery strength on students’ thinking in relation to visuospatial problem solving in chemistry†
Lauren
Baade
a,
Effie
Kartsonaki
a,
Hassan
Khosravi
b and
Gwendolyn A.
Lawrie
*a
aSchool of Chemistry and Molecular Biosciences, The University of Queensland, Brisbane City, QLD 4072, Australia. E-mail: g.lawrie@uq.edu.au
bInstitute for Teaching and Learning Innovation, The University of Queensland, Brisbane City, QLD 4072, Australia
First published on 11th November 2024
Abstract
Effective learning in chemistry education requires students to understand visual representations across multiple conceptual levels. Essential to this process are visuospatial skills which enable students to interpret and manipulate these representations effectively. These abilities allow students to construct mental models that support problem solving and decision making, improving their understanding of complex concepts, for example chemical structures and reactions. The impact of individual differences in mental imagery, such as aphantasia and hyperphantasia, on chemistry students’ spatial thinking when engaging with visual representations is not well understood. This paper presents two exploratory studies that examine how the vividness of mental imagery is related to student outcomes in chemistry-related visuospatial problem solving. The first study quantitatively assessed the performance of first-year university students in tasks requiring complex visual and spatial reasoning within a chemistry context. The second study, involving the same participants, used qualitative interview data to investigate their cognitive strategies with a focus on how their mental imagery impacts their problem-solving approaches. Preliminary results suggest that the vividness of students’ visual mental imagery did not significantly impact their ability to spatially reason with visual representations in chemistry. Our findings also indicate that students with aphantasia may employ alternative strategies that mitigate their lack of visual mental imagery. This paper highlights the need for further research into the diversity of cognitive mechanisms employed by chemistry students of varying mental imagery capabilities.
Introduction
Effective teaching in chemistry utilises visual representations at the macroscopic, submicroscopic, and symbolic levels, guided by Johnstone's triangle (Johnstone, 1982; Cooper and Stowe, 2018). As discussed by Rau (2017), the use of visual representations in science can enhance student learning by making complex concepts easier to understand (Uttal and Doherty, 2008; Schnotz, 2014) and by facilitating instruction using complementary information (Ainsworth, 2006; Ainsworth, 2014). This enables students to deepen their understanding by integrating representations into coherent mental models (Kozma and Russell, 1997; Seufert, 2003; Schnotz, 2005; Schnotz, 2014; Rau, 2017). In learning with representations, students develop and apply the necessary skills that can be regarded as essential for disciplinary practice in chemistry. For example, a chemist who has synthesised a novel compound in the laboratory (macroscopic) may wish to confirm its molecular structure (submicroscopic) by interpreting its NMR spectra (symbolic).Students’ abilities to interpret and translate between representations are often referred to as representational competencies (Kozma and Russell, 1997). Rau (2017) proposes two important subcomponents of these competencies. The first is conceptual representational competencies, which refers to one's ability to notice details, construct an internal representation of them, and link them to existing conceptual knowledge. The second is perceptual representational competencies, which refers to the speed and ease with which one comprehends visuospatial patterns and features, forms an internal representation that depicts them, and then integrates them with prior perceptual knowledge. These two subcomponent skills are contingent on the construction of internal representations which are combined and integrated into cohesive and detailed mental models (Rau, 2017; Rau et al., 2017).
Due to the capacity of mental models to store perceptual details (Rapp, 2005), it is commonly assumed that visual mental imagery is key to their construction. For example, Kozma and Russell (2005) wrote “All chemists have developed the ability to ‘see’ chemistry in their minds in terms of images of molecules and their transformations.” However, the concept of aphantasia, defined in 2015 (Zeman et al., 2015), challenges this notion. Zeman and colleagues defined aphantasia as “a condition of reduced or absent voluntary [visual mental] imagery”. ‘Extreme’ aphantasia, the absence entirely of visual mental imagery, is estimated to be experienced by approximately 1% of the population, and 2–6% experience visual mental imagery that is at best “vague and dim” (Zeman et al., 2020; Dance et al., 2022). At the other end of the spectrum is hyperphantasia, where visual mental imagery is described as akin to seeing externally and comprises around 2.5% of the population (Zeman et al., 2020; Milton et al., 2021). The vividness of visual mental imagery appears to be an inherent trait that is resistant to training (Rademaker and Pearson, 2012). However, it has been observed that mental imagery can be lost in individuals following serious medical events (Zeman et al., 2010; Bumgardner et al., 2021).
Rau proposes that many educators overlook the fact that students may lack essential representational competencies, impeding their learning of novel content (Rau, 2016). There has been limited research into students’ mental imagery abilities in relation to their learning of chemistry, which supports Rau's position. To consider this gap, our study explores the relationship between the strength of students’ mental imagery and their performance in chemistry-related tasks.
Mental imagery
Mental imagery refers to the process of accessing depictive sensory information within the mind (Pearson and Kosslyn, 2015). Visual mental imagery is commonly referred to as seeing with the “mind's eye”. However, mental imagery can also include sounds, smells, tastes, kinaesthetic sensations, and more (Nanay, 2018). Therefore, when discussing imaginative depictive mental experiences in general, we will use the term ‘mental imagery’. When focusing on a specific sensory modality, we will use more precise terms such as ‘visual mental imagery’.In the first known quantitative study on visual mental imagery, Francis Galton asked his scientific friends to describe their imagined breakfast table, focusing on the illumination, definition and colouring (Galton, 1883). To his surprise, most of Galton's respondents denied experiencing visual mental imagery and were sceptical of those who did. Galton's subsequent enquiries into the general population found that many adults and children confirmed that they held visual mental imagery, in contradiction to his initial observations. In his subsequent formal study with 100 men, he found that approximately 5% had little or no visual mental imagery (Galton, 1883), which is consistent with findings from contemporary research (Zeman et al., 2015; Dance et al., 2022).
The absence of visual mental imagery received little attention during the 20th century (see Zeman, 2024 for a discussion). In 2015, Zeman and colleagues devised the term aphantasia. Their inclusion of “reduced” in “a condition of reduced or absent voluntary [visual mental] imagery” is a subject of debate (see Blomkvist and Marks, 2023), but frequently occurs in cognitive psychology literature (e.g.Zeman et al., 2020; Bainbridge et al., 2021; Milton et al., 2021). While we recognise there is a distinction between reduced and involuntary visual mental imagery, compared to its complete absence, for the purposes of this paper we have chosen to conservatively adopt the broader definition of aphantasia. This choice is justified by the prevalence of self-identified aphantasics who describe their condition as a deficiency of voluntary visual mental imagery, rather than an absence entirely (Zeman et al., 2016). Such individuals report experiencing fleeting flashes (Zeman et al., 2016), or dim, dark outlines that quickly dissolve (Dance et al., 2022). Given that our paper focuses on the purposeful use of mental imagery by students in chemistry, we believe it appropriate to use a definition of aphantasia that encompasses both low and involuntary, as well as absent, visual mental imagery abilities. We would additionally like to note that when discussing the range of mental imagery abilities in individuals in a general sense, we will use the terms ‘strength’ and ‘vividness’ interchangeably to describe this spectrum, including those with reduced or absent voluntary mental imagery.
Galton's initial study indicated a potential higher occurrence of aphantasia amongst scientists in comparison to the general population. Similarly, a survey in 2020 showed that of 2000 aphantasics, 32.1% of those who were employed worked in STEM (Zeman et al., 2020). This figure is substantially higher than the 5.7% STEM representation in the U.S. labour force reported in 2022 (U.S. Bureau of Labor Statistics, 2023). Furthermore, the same 2020 survey found that of 200 hyperphantasics, 18.5% of those employed also reported working in STEM. These findings, though likely biased by the voluntary nature of participant involvement, indicate the presence of aphantasics and hyperphantasics in scientific fields.
While research set in chemistry or in STEM education is limited, studies of aphantasia's impact on cognitive skills that are pertinent to representational competencies may offer insight. For example, aphantasics may possess reduced capacity to store multiple different visual representations in their working memory. When asked to redraw real-world scenes in a “Drawing Recall Experiment”, aphantasics were observed to recall fewer objects and colours than non-aphantasics. However, their spatial accuracy was equivalent and they made fewer errors in content (Bainbridge et al., 2021). In contrast, for tasks with simpler visual stimuli, such as striped patterns (Gabor patches) or numbers, no significant difference in the working memory capacity of aphantasics was observed (Keogh et al., 2021). These studies suggest that the complexity and variety of visual elements in a task may influence the working memory performance of individuals with aphantasia. As such, it can be inferred that a functional impact on aphantasic chemistry students’ learning may be more pronounced when using complex visual representations.
Furthermore, aphantasics’ lack of visual mental imagery may contribute to reduced visual attention. For example, Monzel and colleagues (2021) found that aphantasics demonstrated slower response times in their “Spontaneous Use of Visual Imagery Visual Search Task” compared to non-aphantasics. Participants were “explicitly asked to visualize” word cues before choosing between the cue and a distractor, presented in either word or image form. While no differences were seen for words, or in accuracy overall, non-aphantasics were faster at selecting images, which was attributed to a “priming effect” of using visual mental imagery. Additionally, Keogh and Pearson (2021) demonstrated that both aphantasics and non-aphantasics could be primed to make selections by viewing external visual stimuli beforehand. However, only non-aphantasics could be primed using visual mental imagery. These findings collectively indicate that visual mental imagery may enhance one's ability to identify stimuli in real-life situations. Other recent research has provided further evidence for slower visual processing by aphantasics in tasks involving mentally generating or interpreting representations of objects, colours, words, and faces, but not for tasks involving spatial relationships (Liu and Bartolomeo, 2023). These features are all important in representations of concepts and structures in chemistry. Consequently, aphantasic chemistry students may be impacted in their ability to efficiently process and identify the visual representations commonly employed in teaching to illustrate concepts.
Visuospatial ability
In this paper, visuospatial ability refers to a specific aspect of spatial ability that involves perceiving, manipulating, and interpreting spatial relationships among visual representations (Carlisle et al., 2015). Visuospatial ability has long been a topic of interest considered in relation to representational competence in chemistry, and is also considered critical for the development of expertise in other scientific fields (Kell and Lubinski, 2013). Fundamental to visuospatial ability are general spatial skills. Research in the field of psychology shows that aphantasics have self-reported similar spatial ability to non-aphantasics (Keogh and Pearson, 2018; Dawes et al., 2020; Bainbridge et al., 2021) as measured by the “Object and Spatial Imagery Questionnaire” (Blajenkova et al., 2006). In terms of quantitative measures of spatial ability, as previously mentioned, memory and processing of spatial relations have been found to be similar in aphantasics compared to non-aphantasics (Bainbridge et al., 2021; Liu and Bartolomeo, 2023). Furthermore, aphantasics were observed to perform equally well on “Manikin's test” (a mental rotation task) in terms of accuracy (Milton et al., 2021). However, in another study, on two tasks involving object rotation, movement and maintenance, aphantasics showed equal accuracy but longer response times, especially as task complexity increased (Pounder et al., 2022). A further study from 2024 found that aphantasics were slower but more accurate than control participants on two mental rotation tasks (Kay et al., 2024). Collectively, these findings are significant because they suggest that although aphantasics are not inhibited in their ability to correctly solve spatial questions, they may do so more slowly. This raises questions in regard to assessment of learning in chemistry that involves visuospatial abilities, particularly in timed tasks.In summary, aphantasia may impact one's working memory, attentional capacity, and efficiency on spatial tasks. Each of these are important when considering a chemistry student's representational competence and visuospatial thinking. Consider, for example, an exam question involving a complex three-dimensional molecular structure, wherein the student must answer questions relating to its properties and reactivity. The student would need to recall relevant chemical concepts into their working memory. Their understanding must be linked to the molecule's constituents, requiring visual attention. Finally, their capacity to interpret the spatial geometry of the molecule might be required, particularly in such a question where the molecular geometry influences reaction mechanisms and outcomes. If indeed experiencing aphantasia contributes to decreased capacity or efficiency in processing information, it may be necessary for educators to consider accommodation of these capabilities in timed assessment to foster an equitable learning environment.
Spatial problem-solving strategies with chemistry representations
In the field of chemistry education research, the exploration of student approaches to spatial problem solving has garnered significant attention during the 21st century (e.g.Stieff and Raje, 2010; Stieff, 2011; Hegarty et al., 2013; Kiernan et al., 2021). Central to much of this research is the concept of imagistic reasoning, defined by Stieff (2011) as “the process of generating and manipulating perceived analog image-like mental representations for thinking and problem solving”. This concept contrasts to analytic and algorithmic strategies, which involve applying rules and heuristic to spatial and non-spatial information, respectively. Additionally, there are diagrammatic strategies, which refers to modifying existing diagrams to recreate them in a different format (Hegarty et al., 2013). Often however, the latter three strategies are collectively termed ‘analytic’ (Stieff and Raje, 2010; Kiernan et al., 2021), forming a dichotomy with imagistic reasoning in the context of student strategy selection.Researchers have employed various methods to identify the strategies utilised by students, as well as those used by experts, when solving organic chemistry problems. These studies include post hoc self-report data, wherein the subject selects from a list of strategies or describes their own if none match (Stieff et al., 2010; Stieff et al., 2012; Hegarty et al., 2013; Stieff et al., 2014). Additionally, think-aloud protocols in interview settings have been employed (Stieff and Raje, 2008; Stieff and Raje, 2010; Stieff, 2011). Finally, researchers have made assumptions on the type of strategy used by students based on the time taken for subjects to solve questions (Stieff, 2007; Stieff, 2013), the nature of the problems being solved (Vlacholia et al., 2017), and analysing pen-to-paper responses (Kiernan et al., 2021).
Multiple studies have shown that organic chemistry students tend to initially rely on imagistic strategies to solve problems but shift towards the use of more analytic strategies as they gain expertise (Stieff et al., 2010; Stieff et al., 2012; Hegarty et al., 2013; Stieff et al., 2014). The shift was found to be particularly pronounced for students that the authors categorised as female, or as having lower spatial ability (Stieff et al., 2010; Stieff et al., 2012). One study found that the combined instruction of both imagistic and analytic strategies to solve organic chemistry problems eliminated performance differences observed between two groups that the authors categorised as females and males. However, it should be noted that they did not provide a clear explanation for the categorisation. The authors found that training in only one type of strategy resulted in an achievement advantage for only the group categorised as male (Stieff et al., 2014). Therefore, although novices are likely to adopt more analytic strategies as their proficiency increases, continuing to teach imagistic methods (when possible) may still be beneficial.
The role of combined instruction of imagistic and analytic strategies has been further evidenced. Stieff (2011) found that undergraduate organic chemistry students depended on analytic methods, such as drawing, to solve complex questions, but still preferred to use imagistic reasoning for translating between molecular representations. Similarly, chemistry experts were found to use imagistic and analytic strategies in different circumstances, and their strategy choices were influenced by the nature of the task and their individual preferences (Stieff and Raje, 2008; Stieff and Raje, 2010). Hence, whilst experts have been observed to be more inclined than novices to use analytic strategies, they may still prefer at times to use imagistic strategies.
Further research has reported the benefits of teaching imagistic strategies as opposed to only analytic approaches, such as the study by Leutner and colleagues (2009). The researchers found that tenth-graders learning about the dipole nature of water had reduced cognitive load and enhanced understanding when they used mental imagery to learn the concepts, as opposed to drawing them. Additionally, Kiernan and colleagues (2024) observed that senior-level school students were more accurate at determining molecular geometries when they used imagistic reasoning (via speech and gestures) as opposed to using a written analytic method (e.g. VSEPR algorithm).
Other relevant recent literature in chemistry education has increasingly focused on improving instructional scaffolding methods, particularly by encouraging students to consider their metacognitive processes (Graulich et al., 2021; Vo et al., 2022). This approach prompts students to reflect on their thought processes – thinking about thinking – either during or after problem-solving activities. Focusing on metacognitive skills aids students in selecting and adapting cognitive strategies, whether they be imagistic or analytic, to reach solutions rather than mindlessly applying memorised algorithms and formulas. In doing so, instructors empower students to transition from novice to expert problem solvers by enabling them to understand their own cognitive processes and effectively switch between different approaches depending on the demands of the task.
Visual mental imagery and imagistic strategies
The question of whether all chemistry students have the ability to employ imagistic strategies remains unexplored, particularly in relation to aphantasia. One may intuitively assume that aphantasic students are unable to use imagistic strategies, such as “I tend to imagine the molecule in 3D and rotate it ‘in my head’” (Stieff et al., 2012), due to their inability to readily form visual mental imagery. Contrary to this assumption, however, several studies (Keogh and Pearson, 2018; Dawes et al., 2020; Bainbridge et al., 2021) have reported that aphantasics and non-aphantasics score similarly on the spatial component of the “Object and Spatial Imagery Questionnaire” (OSIQ) (Blajenkova et al., 2006). This questionnaire rates responses from 1 to 5 on items such as “I can easily imagine and mentally rotate 3-dimensional geometric figures”. Keogh and Pearson (2018) found these results unexpected, noting that aphantasic participants consistently gave low scores for the object component – a traditional measure of visual mental imagery – yet did not score similarly for the spatial component. This suggests that the inability to readily form visual mental images does not necessarily inhibit the use of imagistic strategies for problem solving, challenging the notion that imagistic reasoning requires “image-like mental representations” (Stieff, 2011). However, if imagistic reasoning does require visual mental imagery, the similar scores of aphantasics and non-aphantasics on the OSIQ's spatial component may indicate that aphantasics can achieve the same visuospatial outcomes as non-aphantasics, but through non-imagistic mechanisms. This would suggest that the wording of the OSIQ is not precise enough to distinguish between the mental processes of aphantasics and non-aphantasics.Furthermore, Blazhenkova (2016) introduced the concept of “visual-spatial imagers” who excel in visuospatial tasks by using mental imagery to represent spatial relations and transformations. This contrasts with “visual-object imagers” who use mental imagery for vivid and colourful images, and excel in tasks that require “object visualization”. Both types, however, are described using terms that imply visual mental imagery. Later, Blazhenkova and Pechenkova (2019) proposed that aphantasics “may preserve intact spatial imagery or even excel in spatial mental visualization”. They cite the case study of “patient MX” (Zeman et al., 2020), who lost the ability to generate visual mental imagery but still performed well on a mental rotation task. While Blazhenkova and Pechenkova interpret this as evidence of intact visual-spatial imagery, the original study explained that MX used a “perceptual matching strategy” which did not rely on visual mental imagery. This raises questions about their definition of “visual-spatial imagery” and suggests more clarity is needed. Can individuals with little or no visual mental imagery perform visuospatial tasks in the same way as those without aphantasia, or must they rely on alternative strategies like perceptual matching? If so, is it appropriate for instructors to teach imagistic strategies to students if some students, like those with aphantasia, cannot use them?
A study by Kay and colleagues (2024) offers additional insight. They found that aphantasics, while slower, were more accurate in solving mental rotation questions (Manikin's test) compared to non-aphantasic controls. In a post-test questionnaire, 24% of aphantasics agreed or strongly agreed with a statement indicating the use of an imagistic strategy. For example, one aphantasic participant described attempting to “imagine rotating the man so that he was facing away from me, though I could only rotate him for a second or so since I can’t ‘see’ it”. This suggests that aphantasics might use imagistic strategies in ways that differ from typical visual mental imagery processes. Additionally, the authors’ data analysis showed that strategy choice best explained differences in response times, with non-imagistic strategies resulting in slower performance. However, visual mental imagery ability best explained differences in accuracy, with aphantasics outperforming controls. Thus, a question remains: if Stieff's (2011) description of imagistic strategies requiring “image-like mental representations” holds true, we would expect differences in aphantasic students’ approaches and performances on visuospatial chemistry tasks to become evident.
Purdue spatial visualisation test (PSVT)
A test that has been widely adopted to measure students’ visuospatial abilities in the context of chemistry education, amongst others, is the “Purdue Spatial Visualisation Test” (PSVT) (Maeda and Yoon, 2013). Originally created by Guay in 1976, this test has seen various iterations, the most notable of which is titled the “PSVT:Rotations” (PSVT:R) (Yoon, 2011). When solving PSVT and PSVT:R questions, participants view a complex three-dimensional object's spatial transformation, then choose from five options which one is a second, different object transformed in the same manner.Performance on the PSVT has been found to positively correlate with chemistry proficiency in undergraduate students (Carter et al., 1987; Pribyl and Bodner, 1987; Yang et al., 2003). This suggests that the PSVT may capture a skill that occurs concurrently with achieving success in chemistry learning. Furthermore, expert chemists were found to outperform specialists in computer science and education on the PSVT:R (Hall et al., 2021). These chemists used discipline-specific language such as “rotational axes” to describe their problem-solving process, which provides further evidence of the cognitive link between PSVT questions and skills used in chemistry. Additionally, the relationship between PSVT performance and chemistry ability may be bidirectional. Undergraduate students who enrolled for the first time in an organic chemistry course were tested on the PSVT pre- and post-instruction (Hornbuckle et al., 2014). Students initially categorised as “below average” on the PSVT showed substantial improvement on the PSVT at the end of the chemistry course, whereas “above average” students did not change in their performance. Similarly, a study by Cole and colleagues (2020) found that middle school students’ PSVT scores moderately correlated with their understanding of matter conservation, and both their PSVT scores and this correlation increased following matter conservation-targeted instruction. To summarise, the correlation between PSVT performance and chemistry proficiency as evidenced by these studies emphasises the test's potential as an indicator of skills relevant to success in chemistry.
The specific strategies employed by chemistry students to solve PSVT questions is unexplored to date. In a study of students within a third-year chemistry course focusing on spectroscopy, Southam and Lewis (2013) proposed that students who performed well on the PSVT likely used imagistic strategies. Conversely, students who scored lower on the PSVT were presumed to use alternative strategies. The authors reference Willis and colleagues (1979), who studied the brain activity of high school students as they completed various tasks, including the PSVT. They observed different EEG patterns between high and low PSVT performers and concluded that high performers were utilising “imaginal movement” (Willis et al., 1979). This inference was made without substantial evidence linking specific EEG patterns to the use of visual mental imagery, nor by explaining why such an imagistic approach would be more prevalent among high scorers. In fact, a review by Bartlett (2023) summarised that there is no evidence to date of a relationship between strategy and performance on the PSVT, nor are there any conclusive studies on what strategies exist other than mental rotation. Nonetheless, the study by Southam and Lewis found that regardless of their initial visuospatial ability (as measured by their PSVT scores), students performed equally well in point group theory assessments after instruction involving both imagistic and analytic strategies (Southam and Lewis, 2013). Although lacking in a control group, the finding suggests that the combined instruction of imagistic and analytic strategies can minimise the PSVT's predictive power on chemistry achievement.
A review by Stieff and Uttal (2015) found insufficient evidence to support that spatial training can improve student STEM success. This could indicate that training specifically designed to improve scores on the PSVT may not lead to significant improvements in chemistry performance. It has been suggested that PSVT might assess abilities beyond mental rotation, such as interpreting complex isometric projections (Bartlett, 2023), which have similarities to three-dimensional molecule representations and may explain the observed correlation between PSVT performance and chemistry skills. Furthermore, Stieff and colleagues (2018) argue that spatial thinking in chemistry depends more on students’ representational competence than on their inherent spatial ability. Ultimately, the precise mechanisms underlying the relationship between PSVT performance and chemistry proficiency remain unclear. Nevertheless, value remains in the PSVT as not only a tool to measure visuospatial skills, but also to provide insight into the cognitive processes that underpin learning in chemistry. Our research aims to explore the impact of mental imagery on PSVT performance, and investigate connections between cognitive approaches and their effectiveness in developing the visuospatial skills vital for success in chemistry.
Chunking and mental imagery
As students develop expertise in processing visual representations in chemistry, they utilise cognitive strategies such as chunking. Chunking theory from cognitive psychology suggests that as expertise increases, individuals can incorporate larger amounts of detailed information into mentally encoded units known as ‘chunks’ (Miller, 1956; Chase and Simon, 1973). This process reduces the demand on working memory and frees up additional cognitive resources. Stieff and colleagues (2020) explored this concept by examining visual chunking – the grouping of visual information specifically – and found that it improved the ability of both chemistry-novice and chemistry-naïve undergraduates to detect changes in molecular representations. This suggests that visual chunking may enhance students’ abilities to encode and recall complex spatial information – an important skill in chemistry. However, while chunking and mental imagery have been explored as related concepts in areas such as chess (Saariluoma and Kalakoski, 1997; Waters and Gobet, 2008) and sports (Schack et al., 2014), no studies have specifically examined the link between the vividness of mental imagery and chunking effectiveness. Therefore, recognising the value of chunking in the development of chemistry expertise, this study aims to investigate associations between mental imagery vividness and students’ abilities to chunk information in the context of chemistry-related tasks.Measuring mental imagery abilities
The strength of an individual's mental imagery must be measured to compare how this trait impacts on their cognitive function. Neuroimaging research seeking biological markers to confirm the varying strengths of visual mental imagery among individuals has found correlations between self-assessments of visual mental imagery vividness and activity in various regions of the brain (Cui et al., 2007; Bergmann et al., 2016; Fulford et al., 2018; Milton et al., 2021). Such self-assessments are conventionally completed using a self-report questionnaire. The “Vividness of Visual Imagery Questionnaire” (VVIQ) (Marks, 1973) is the most well-known and widely used, however is limited because it only assesses an individual's strength of visual mental imagery. As pointed out by Andrade and colleagues (2014), mental imagery can occur in sensory modalities other than vision – for example, an “inner dialogue” may occur through sound. Empirical data on the prevalence of aphantasia in other modalities is currently lacking. It is known, however, that an individual's mental imagery may be equally weak or vivid across all modalities, or may be strong in some but faint in others (Zeman et al., 2020; Dawes et al., 2024). The lack of a suitable questionnaire to assess multi-modal mental imagery led to the creation of the “Plymouth Sensory Imagery Questionnaire” (Psi-Q), which measures vividness of mental imagery over a range of sensory modalities – vision, sound, smell, taste, touch, bodily sensation, and emotional feeling (Andrade et al., 2014). The Psi-Q has been used in medical research (Mitrenga et al., 2019; Tomasino et al., 2022), but not yet in educational research. The capacity of the Psi-Q to measure mental imagery across a range of modalities, not just visual, makes it an appealing research tool. This paper has established that visual mental imagery affects cognitive task performance, however the role of mental imagery in other senses has yet to be explored in this context.Another instrument of interest is the “Vividness of Object and Spatial Imagery” (VOSI) questionnaire (Blazhenkova, 2016). As mentioned earlier, the VOSI focuses on only visual mental imagery but distinguishes it into two subcomponents – visual-object mental imagery and visual-spatial mental imagery. The former refers to the conventional understandings of visual mental imagery, referring to objects, scenes, shapes, colours, etc. The latter, however, refers to the imagining of “spatial relations and movements of objects and their parts, and spatial transformations”. When applied to university students, the authors found the two mental imagery types to be distinct but correlated, with visual-object mental imagery scores found to be related to object recognition skills, artistic ability, and inversely to self-assessment of scientific ability. Conversely, visual-spatial mental imagery was positively correlated with spatial ability (as measured by the “Paper Folding test”) and self-assessment of scientific ability. In a later article, co-authored by the creator of the VOSI, the researchers introduced the concept of “spatial aphantasia”, a separate condition to what they call “object aphantasia” (Blazhenkova and Pechenkova, 2019). The authors advocate for further investigation into spatial aphantasia, particularly its cognitive implications, related deficits, and compensatory strategies. As explained above, it remains unclear whether deficits in “visual-spatial imagery” have the same implications for aphantasic and non-aphantasic individuals.
Rationale
The present study aims to investigate the contribution of mental imagery strength on students’ thinking in relation to visuospatial problem solving in chemistry. The overarching focus of this study is the development of quantitative and qualitative design protocols that can capture and explore any relationships between mental imagery vividness and the processing of visuospatial chemistry representations. Two studies have been conducted to address the following research questions:1. To what extent does the strength of students’ mental imagery impact their thinking when dealing with representations in chemistry-related visuospatial tasks? (RQ 1).
2. Does aphantasia impact on students’ abilities to employ imagistic strategies in chemistry problem solving? (RQ 2).
These studies were designed as an exploration into students’ mental imagery abilities in relation to their thinking of chemistry. In the first study, quantitative data were collected from first-year university chemistry students to test a hypothesis that students’ mental imagery abilities relate to their performance in chemistry-related tasks. In the second study, interviews were conducted, involving the same sample of participants, to gain insights into their cognitive processes and the extent to which they adopt imagistic strategies. The procedures used in each study were approved by the institutional ethical review committee (ID# 2021/HE002251).
Research method
This study uses a sequential mixed methods research design (Leech and Onwuegbuzie, 2009). The first study was designed as exploratory quantitative data collection, aiming to develop a chemistry-discipline-specific diagnostic instrument that can investigate relationships between students’ mental imagery vividness and their capacity for cognitive skills required to learn chemistry. In summary, first-year undergraduate chemistry students completed several chemistry-related tasks, and their performances were compared to their self-assessments of mental imagery strength. A subset of this data is presented here, forming part of a larger doctoral research study – for further details please contact the authors. The second study involved interviews with the participants from the first study to gain qualitative insights into their mental processes. Details of both studies are described below.Informed by the review of literature, it was anticipated that distinct differences in performance between students with high and low mental imagery abilities might be observed. As described earlier, prior research had found differences between aphantasics and non-aphantasics in working memory, attentional capacity, and efficiency on spatial tasks. We hypothesised that students with low visual mental imagery would generally perform worse on chemistry-related tasks in terms of speed and/or accuracy. Furthermore, we anticipated that there would be differences in the processes of thinking employed by the participants in relation to their mental imagery strengths.
Trialling diagnostic tools (quantitative)
![[thin space (1/6-em)]](https://www.rsc.org/images/entities/char_2009.gif)
000 students. Undergraduates enrolled in a second-semester, first-year chemistry course (n = 286) were recruited via an email sent through the course learning management system. Convenience sampling based on the criteria of being currently enrolled in the course was used. Despite efforts to incentivise participation (such as an entry into a draw to win a $100 gift card) the response rate was poor. Additionally, recruitment was conducted midway through the teaching semester to ensure that all relevant key concepts required to complete the activities in the survey had been taught. Hence, this timing may have contributed to the small sample size, as students likely faced competing academic priorities and time constraints during this period. Ultimately, eighteen (n = 18) students provided informed consent and participated in this study.
 | ||
| Fig. 1 Overview of the survey completed by participants in the first study. For information not included in this paper, please contact the authors. | ||
 | ||
| Fig. 2 Examples of a text representation (left) and an image representation (right) of a carboxylic acid, used in the SUVI-FG. | ||
For the SUVI-LO, participants completed 28 trials (14 text, 14 images), and for the SUVI-FG they completed 30 trials (15 text, 15 images). Participants were familiarised with the images used in both tasks prior to their commencement. Instructions guided participants to “visualise the image of the cue in your mind to the best of your ability”. They were then shown the name of a lab object (SUVI-LO, e.g. “beaker”) or a functional group (SUVI-FG, e.g. “carboxylic acid”), followed by an image of a crosshatch symbol (+) as a pause, and then two representations involving either a text or an image format of the cue and a decoy. Participants then used the left or right arrow key to select which representation corresponded to the cue. Fig. 3 provides examples of a text and an image trial for the SUVI-LO and SUVI-FG. Both a response time and the correctness of response for each trial was recorded. Additionally, participants completed three warm-up trials prior to data collection, and these indicated correctness as to ensure the task was understood, however correctness was not indicated during the main trials.
 | ||
| Fig. 3 Examples of text and image trials from the SUVI-LO (left) and SUVI-FG (right). | ||
To validate the two SUVI adaptations prior to the main data collection, three interviews were conducted with students from past cohorts of the course to gain feedback on the tasks. This ensured that the tasks’ instructions were comprehensible, and that the choice of representations used were interpretable and distinct from each other.
The first five questions presented to students involved objects rotated once by 90 degrees. The following set of three questions involved two 90-degree rotations. The final set of two questions involved three 90-degree rotations. See Fig. 4 for an example of one of the questions that involved two 90-degree rotations. Instructions were provided to the participants that were identical to instructions for the PSVT:R, and there were no warm-up trials. Accuracy and time taken by the participants for each of the ten questions were recorded.
 | ||
| Fig. 4 An example PSVT:R_A question that involves two 90-degree rotations. | ||
With regards to the tasks (SUVI-LO, SUVI-FG and PSVT:R_A), accuracy and response times were collected through PsyToolkit. To analyse the time data, only correct answers were included and incorrect answers were excluded. Filtering for correct answers provided a clearer measure of cognitive processing speed by eliminating variability potentially introduced by, for example, guessing, misunderstandings, or lapses in attention (De Boeck and Jeon, 2019). In addition, we used reciprocal time values (1/time, speed) to analyse the time data. This is a common transformation applied in cognitive psychology reaction time research (Whelan, 2008). The transformation was beneficial as it normalised the data, elucidated statistically significant relationships, and minimised the effects of outliers. With regards to outliers, although there are various methods for determining cut-off limits (Ratcliff, 1993), in general the cautionary path of including all data was taken to avoid erroneously removing legitimate times. Using reciprocal time data results in longer reaction times contributing less statistical impact, and thus minimised the effect of potential outliers, without removing them entirely. However, there were two exceptions made where data clearly warranted exclusion. One participant's data was excluded from analysis of the PSVT:R_A questions due to evidence of disengagement, indicated by their unusually fast response times combined with a consistent lack of correct answers. Additionally, a separate participant was excluded from analyses involving the Psi-Q, because they consistently recorded the exact same response across all items. This was in contradiction to their responses on the VOSI, which appeared normal, leading to the decision to disregard their Psi-Q scores due to potential unreliability.
We used linear mixed models (LMMs) to analyse time data and generalised linear mixed models (GLMMs) with a binomial distribution to analyse correctness data, both in comparison to participants’ questionnaire scores. LMMs were chosen for their ability to incorporate random effects, which capture individual variability that is not explained by fixed effects (e.g. questionnaire scores) (Pinheiro and Bates, 2000). This approach is particularly beneficial for our analysis as the data involves multiple observations collected per participant. Additionally, LMMs are well-suited to handle datasets with imbalanced data distributions, as they can account for variability across different levels of the data (Gałecki et al., 2013). This characteristic allows LMMs to mitigate the influence of any particular subgroup of participants on the overall model estimates. Although we did not explicitly categorise participants by their mental imagery strength, the rarity of aphantasia may lead to skewed distributions of scores. LMMs help address this by allowing for more flexible modelling of the data. Nonetheless, we recognise that the fixed effects in our models may predominantly reflect trends from higher-scoring participants due to greater availability of data. Additionally, the small sample size may increase the likelihood of Type II errors, preventing the detection of significant effects, and may produce less reliable variance estimates, resulting in wider confidence intervals and reduced statistical power. This could undermine the validity of hypothesis testing by making it more difficult to draw definitive conclusions. However, by including a random effect for each participant we account for additional variability, which improves the robustness of the models.
Another advantage of using LMMs is that independent variables do not need to follow a distribution, normal or otherwise, which is particularly beneficial to our analysis given the anticipated imbalanced distributions of mental imagery strength. Instead, the residuals and random effects are assumed to be normally distributed (Gałecki et al., 2013). We verified these assumptions through Quantile–Quantile (Q–Q) plots, Shapiro-Wilk tests, and by assessing homoscedasticity and linearity with plots of fitted values against residuals. We did this to try to ensure the integrity and validity of our model outcomes, particularly in light of the challenges caused by imbalanced data distributions and small overall sample size.
For the GLMMs, we verified that our models accurately represent underlying data structures and provide reliable predictions by assessing model performance using the Area Under the Curve (AUC) from the Receiver Operating Characteristic (ROC) curve. Additionally, we confirmed the linearity in the logit transformation of predicted probabilities and checked for overdispersion to ensure the appropriateness of the binomial model.
For both LMMs and GLMMs, we assessed model fit by calculating R-squared values, Akaike Information Criterion (AIC), Bayesian Information Criterion (BIC), and log-likelihood. For models involving more than one fixed effect, we further examined significant findings by calculating estimated marginal means and conducting pairwise comparisons.
Finally, to address any concerns about multicollinearity among the subcomponents of the mental imagery questionnaires, we tested for correlations and calculated Variance Inflation Factors (VIF).
Exploring participants’ cognitive strategies (qualitative)
Part A. The participants were first asked what university degree they were currently completing, the last chemistry course they had completed, and whether they found their chemistry knowledge useful in their studies if they were not a chemistry major. They were also asked what they remembered of the survey, what they understood by mental imagery, and whether they recalled any instances of when they were required “to visualise something in [their] mind” throughout their learning. For the purposes of the interview and convenience of communication, we conflated ‘visual mental imagery’ and ‘mental visualisation’, explaining to the participants that we would use the terms synonymously.
Part B. The participants were asked what was meant by the term ‘functional group’, and any necessary clarifications were made. Next, they were asked to imagine a carboxylic acid and to indicate to the interviewer once they had finished. They were then asked to describe the steps that had just occurred mentally, then draw what they imagined. This was then repeated with an ester group. The participants were asked if they provided consent to being timed for how long they took to imagine the carboxylic acid and ester, of which all agreed. The times were measured to millisecond accuracy post interview using audio software Audacity. An important component of the SUVI-FG in the quantitative study was the testing of a priming effect from pre-emptively constructing a visual mental image of a functional group before selecting between it and a distractor functional group. We aimed to explore whether the participants’ mental imagery strength influenced the way they imagined the functional groups, particularly in relation to their efficiency in chunking information, and how long it took them.
Part C. The participants completed three PSVT:R_A questions they had previously encountered during the quantitative tasks. Each of the three questions involved either one, two or three sets of 90-degree rotations of an object, and as such were used to observe if there were differences in the participants’ approaches depending on the complexity of the question. Part C was informed by the findings of the quantitative study where differences in speed and accuracy between the different numbered 90-degree rotation questions were observed, as well as relationships between several modalities of mental imagery strength and time taken. As such, the participants were asked to describe their mental processes while they solved the three questions. They were then asked questions relevant to the PSVT:R_A, such as how easy or hard they find such questions, and whether they think doing such types of questions is a necessary skill to have for doing or learning chemistry.
Part D. The participants were asked to explain how they would give instructions to another student for imagining a carboxylic acid, whether they had any thoughts on the interview, particularly about mental imagery, and any necessary clarifications were made.
For the quantitative interpretation of coding outcomes, we used linear models (as opposed to linear mixed models) due to the lack of repeated measures, removing necessitation for random effects. The process for validating statistical assumptions and evaluating model fit remained otherwise the same. It is important to note that the sample size decreased further in this study, from 18 to 7 participants, which additionally limits the generalisability of the results. Therefore, the findings should be regarded as preliminary and interpreted with caution due to these limitations.
Although considerable time had passed, we decided to reuse the participants’ questionnaire scores from the first data collection. This decision was made based on a general lack of literature evidence to support that a person's mental imagery vividness changes over time. One study observed a decline in the vividness of visual mental imagery among 2252 participants from adolescence to middle age (Gulyás et al., 2022). However, the study acknowledges it was neither representative nor longitudinal, and it did not assess changes in mental imagery vividness over a one-year period. In contrast, another study reported no significant changes in self-reported visual mental imagery vividness after engaging in daily hour-long training sessions over five days (Rademaker and Pearson, 2012).
Results
Developing diagnostic instruments to explore mental imagery and visuospatial thinking in chemistry
The findings from 18 participants who volunteered to be part of this study as a convenience sample from a larger cohort of 286 students are considered here. This small sample size (n = 18) should be considered when interpreting these results, because it is likely to affect both the generalisability of the findings and the statistical power of analyses. Furthermore, this self-selected sample may introduce bias and limit the validity of the conclusions that can be drawn. Despite these limitations, the findings provide preliminary evidence and have served as a basis to inform subsequent research with larger cohorts. Additionally, we discussed in the previous section how we employed linear mixed models (LMMs) and generalised linear mixed models (GLMMs) to help manage challenges arising from our small and uneven sample sizes.| Questionnaire | Component | Mean | SD |
|---|---|---|---|
| VOSI (out of 5) | Visual-object | 3.7 | 1.1 |
| Visual-spatial | 2.9 | 0.9 | |
| Psi-Q (out of 10) | Vision | 7.4 | 2.4 |
| Sound | 6.3 | 2.5 | |
| Smell | 5.4 | 2.9 | |
| Taste | 6.9 | 2.5 | |
| Touch | 7.4 | 2.4 | |
| Bodily sensation | 7.2 | 2.2 | |
| Emotional feeling | 6.8 | 2.4 | |
 | ||
| Fig. 5 Boxplots of participants’ scores across the subcomponents of the VOSI and Psi-Q, from the first study. | ||
We assessed participants’ mental imagery strength using arithmetic mean scores of the VOSI (Visual-Object and Visual-Spatial) and Psi-Q (Vision, Sound, Smell, Taste, Touch, Bodily sensation and Emotional feeling) questionnaires. We then analysed whether the scores, intended to represent the underlying construct of vividness of mental imagery, were associated with variations in participants’ response times across the four representation/format combinations, using linear mixed models. In contrast to the findings of Monzel and colleagues (2021), our analysis showed no significant effects of VOSI Visual-Object or Psi-Q Vision scores on response times for images of lab objects or functional groups. The only significant finding was the relationship between VOSI Visual-Spatial scores and the selection of images of functional groups, with higher scores associated with small increases in response times (p = 0.022). See ESI,† S7 for further details and a visualisation of the model and data.
With regards to accuracy, higher VOSI Visual-Spatial scores were associated with an improved likelihood of correctly identifying images of functional groups by an odds ratio of 2.8 (p = 0.045). This means that for each one-unit increase in VOSI Visual-Spatial score (between 1 and 5), the likelihood of correct identification increased by 180%. Further details can be found in ESI,† S8. Additionally, we found that as VOSI Visual-Spatial scores increased by one unit, there was a 59% decrease in the likelihood of accurately selecting functional group texts, indicated by an odds ratio of 0.41 (p = 0.042). Further details can be found in ESI,† S9. Finally, we found that each one-unit increase in Psi-Q Bodily sensation score was associated with a 35% increase in the likelihood of accurately selecting images of functional groups, as indicated by an odds ratio of 1.35 (p = 0.046). Further details can be found in ESI,† S10.
We compared the participants’ speed at accurately solving PSVT:R_A questions to their mental imagery scores. No relationships were found for questions involving one 90-degree rotation. However, for questions that involved grouped two and three 90-degree rotations, several relationships were found. Longer solving times were associated with higher VOSI Visual-Spatial scores (p = 0.0059, see ESI,† S13 for details and a graph), higher Psi-Q Sound scores (p = 0.0023, see ESI,† S14) and higher Psi-Q Bodily sensation scores (p = 0.0098, see ESI,† S15).
To address concerns about multicollinearity, we tested for correlations between the participants’ VOSI Visual-Spatial, Psi-Q Sound, and Psi-Q Bodily sensation scores. Given the data was not normally distributed, we calculated Spearman correlations. The pairwise correlations ranged from 0.49 to 0.66, indicating a moderate degree of shared variance. Additionally, we included all three variables into one model and calculated the Variance Inflation Factor (VIF), which ranged from 2.05 to 3.05. These VIF values indicate no significant multicollinearity (VIF < 10). Therefore, while the variables are moderately correlated, multicollinearity is not strong enough to impact the interpretation of the individual models.
Investigating relationships between mental imagery scores and accuracy on PSVT:R_A questions yielded no statistically significant results.
Exploring students’ strategies in tasks that may utilise mental imagery (interviews)
Findings from the first part of our study, involving 18 participants self-selected from a cohort of 286 students, revealed several potential quantitative relationships between measures of chemistry-related skills and students’ mental imagery strength. Questions about the specific cognitive processes underlying these relationships prompted us to conduct follow-up interviews with seven of these 18 participants approximately 12 months later. This focused approach allowed for a comprehensive exploration of the cognitive strategies employed by students with different levels of mental imagery abilities, such as aphantasia. However, the sample size (n = 7) limits the generalisability to the initial cohort or similar populations. Nonetheless, the interviews provided insights into the cognitive processes employed by chemistry students during problem solving.We anticipated that the modes of thinking and strategies employed to solve chemistry-related tasks by students with low mental imagery ability would differ greatly from students with high mental imagery ability. This prediction was based on the initial statistically significant relationships found between several modalities of mental imagery strength and performance on the visuospatial tasks. However, it should be noted that only one student with low mental imagery ability was included in the interview participants, and hence our findings cannot be generalised to all students with low mental imagery abilities. Nevertheless, our findings provide insights into how such students might approach chemistry-related tasks. For a table with the questionnaire scores of the seven participants, see ESI,† S3.
The primary research question informing the interviews was: How does aphantasia impact on students’ abilities to employ imagistic strategies in chemistry problem solving?
| Theme | Number of instances | Example reference |
|---|---|---|
| Molecular structure | 2 | “We had to draw out […] the shape of molecules from like a chemical formula, and I thought well, first of all you have to be able to sort of visualise that. It will be sort of difficult to draw it if you can’t visualise, I suppose.” (Jamie) |
| Mechanisms | 2 | “It's almost impossible to try and comprehend […] especially organic chemistry, without picturing the molecules in your head and what they’re doing and how they’re interacting to understand those mechanisms.” (Bailey) |
| Biochemistry and cellular processes | 3 | “Receptors and how they interact with the ligands… the pathways for certain processes in cells.” (Jordan) |
| “How enzymes are interacting with substrates.” (Sam) | ||
| “Receptor interactions, virus interactions.” (Bailey) | ||
Taylor, who scored as having low visual mental imagery ability (VOSI Visual-Object score: 1.3/5), said, “Generally pretty easy, but it kind of depends I guess, ‘cause often those things kind of have a pattern to them. [First-year chemistry] I found pretty easy and it got more harder when we got into harder visualisations in [second-year chemistry].” Bailey, who has high visual mental imagery ability, and Taylor both reported mixed easiness. This suggests that vividness of visual mental imagery may be independent of the difficulty experienced in constructing visual mental images. In further support of this notion, another participant who scored indicating very vivid visual mental imagery felt that they were worse than their peers at actually using it meaningfully – Jamie (VOSI Visual-Object score: 4.8/5) described, “I find it takes me a long time. [….] Sometimes it takes like multiple attempts, and sometimes I have to sort of try again. I don’t think I can do it as quickly as some of my peers seem to be able to, but like, I don’t really know I guess.” Conversely, Sam (VOSI Visual-Object score: 4.4/5) who scored lower than Jamie said they “visualise things a lot”, and when imagining complicated things they “find that pretty easy. […] If I’m reading a book or something, I will be imagining what's happening.”
Rather than referring to increases in complexity or abstractness as being associated with more difficulty in using visual mental imagery, Jordan (VOSI Visual-Object score: 3.9/5) explained, “I think I’m pretty good, like it's not difficult but sometimes I realise that if I learn something the wrong way and then it's corrected after a while, it's harder to visualise that thing. […] Like the structures kind of mix together and its annoying.”
| Participant | Time (s) | Code | Number of steps |
|---|---|---|---|
| Bailey | 1.5 | (1) All at once | 1 |
| Sam | 2.6 | (1) Carbonyl (oxygen) | 2 |
| (2) Alcohol | |||
| Pat | 2.5 | (1) Carbonyl | 3 |
| (2) Alcohol | |||
| (3) Chain | |||
| Jamie | 4.2 | (1) Chain | |
| (2) Carbonyl (oxygen) | |||
| (3) Alcohol | |||
| Taylor | 5.1 | (1) Carbonyl (carbon) | |
| (2) Carbonyl (oxygen) | |||
| (3) Alcohol | |||
| Alex | 3.5 | (1) –COOH | 4 |
| (2) Two lines | |||
| (3) Alcohol | |||
| (4) Carbonyl | |||
| Jordan | 3.6 | (1) Line | |
| (2) Carbonyl (carbon) | |||
| (3) Carbonyl (oxygen) | |||
| (4) Alcohol | |||
Additionally, time was negatively associated with the participants’ Psi-Q Vision scores (p = 0.037), indicating a relationship between more vivid visual mental imagery and faster times to imagine the carboxylic acid. See ESI,† S18 for more details. Psi-Q Vision score (nor any other modality of mental imagery score) was not found to be related to number of steps.
| Strategy | Number of participants who used it | Explanation | Example quote/s | ||
|---|---|---|---|---|---|
| Q1 | Q2 | Q3 | |||
| Mental rotation | 4 | 6 | 3 | Participant mentally visualises the transformation of the original object, then replicates this transformation for the second object. Their focus may be on the object on a whole, or on specific features. Transformations may be approximate and/or stepwise. | • “First of all, I visualised it sort of tipping over to get in like the right sort of position, […] and then I’ve sort of rotated it in my head and knew it should end up like that [second] image, and then I just did the same thing with the second [object].” (Jamie, VOSI Visual-Object score: 4.8/5) |
| • “The actual act of the motion is kind of blurry. I feel like my brain's filling in something random […] like if you take a photo with someone like running in it, and it's just kind of like that sort of blurry thing. But you know they’re running, they’re not just like a blurry blob.” (Pat, VOSI Visual-Object score: 4.1/5) | |||||
| • “I’ll pick a feature and that’ll become almost like a head, and that's (gesturing) like a head and a tail. […] I’d imagine shape one step one, shape two step one, and then the first shape step two, and then the second shape step two.” (Alex, VOSI Visual-Object score: 3.8/5) | |||||
| Projection | 1 | 0 | 1 | Participant focuses on a specific feature of the original object, identifies its new position, and then determines where the corresponding feature on the second object would be. | • “I could see that it had been turned around, and then I see it's like as if the top panel then changes in accordance with being able to now see the back panel […] I followed the same process that I applied to the top one and went looking for the back of the shape.” (Bailey, VOSI Visual-Object score: 4.9/5) |
| • “I think I cheated in that question, but I’m done […] I saw a U on the bottom of [the original object] and I just flipped it around, and then I looked at all the answers, and saw there was one that had a square on the [top] […] I wasn’t certain on the rotation of the rest of the [object], I only saw the square on top […] I just understood that it had gone from bottom to top, which I guess was flipping – I didn’t imagine it flipping.” (Sam, VOSI Visual-Object score: 4.4/5) | |||||
| Verbal | 1 | 1 | 1 | Participant mentally encodes a sequence of verbal instructions describing the transformation of the original object. Mental descriptions of each of the possible answers’ transformations are then compared to the original until the sequences match. | • “Looking at each part of it and then mapping it to the second one and going OK, for A to get to B, it would have to move like this, and move like that. And then looking at […] each of the options and going OK, for it to get to that one, it’d have to move like that. That's a different movement for it to get to that one.” (Taylor, VOSI Visual-Object score: 1.3/5) |
| Multiple | 1 | 0 | 2 | Participant uses more than one of the three other strategies. | • Bailey (VOSI Visual-Object score: 4.9/5) described using mental rotation to see that the original object had been flipped, but then said, “then it needs to be rotated to the back and I’m not seeing the rotation kind-of direction”. They gave up on determining what the final movement was and moved on to looking for the answer, saying, “I realised that I didn’t have to expend the effort of trying to figure out directions and I could see the answer based on the information I’d already determined from the top, which was that the bottom comes to the top (projection).” |
Additionally, we found no relationship between strategy choice and complexity of question (one, two or three 90-degree rotations, including testing with two and three rotation questions as grouped and ungrouped).
We would like to note that several participants attempted to verbally describe their mental processes simultaneous to solving the questions, while the others solved the question first and then described their process afterwards. For this reason, we did not time how long it took them to solve the questions.
Case studies of opposite ends of the visual imagery strength spectrum
| Questionnaire | Component | Case study 1 scores | Case study 2 scores |
|---|---|---|---|
| VOSI (out of 5) | Visual-object | 1.2 | 4.9 |
| Visual-spatial | 1.3 | 3.6 | |
| Psi-Q (out of 10) | Vision | 1.8 | 9.6 |
| Sound | 5.2 | 10 | |
| Smell | 0.4 | 8.6 | |
| Taste | 6.4 | 6.8 | |
| Touch | 6.6 | 6.8 | |
| Bodily sensation | 7.4 | 8.4 | |
| Emotional feeling | 1.2 | 7 | |
At points during the interview, they indicated that they experienced visual mental imagery, but it's “not extremely vivid […] I don’t do it very well”, and never mentioned actively experiencing visual mental images throughout the interview. They instead persistently referred to “impressions”, which they described as “more of like a knowing it's there rather than being able to see it kind of thing, like you kind of think of it, but don’t actually see it as such. […] Kind of like when you know someone is standing behind you, […] you’re not looking at them, but you know that there's someone standing there sort of thing, like you know it's there, but you can’t exactly see or hear or feel it as much as directly.” For these reasons, we determined Taylor to experience aphantasia in terms of reduced voluntary visual mental imagery, rather than its absence entirely.
Taylor said that for them, imagining a carboxylic acid was “a running description of it in words sort of thing, and […] kind of impressions, knowledge of, in space, where it would be. […] Like, that would be here, and that would be there, and that would be there kind of thing.” They were asked if they could imagine anything without describing it in words and replied, “Not really, no.” When asked whether knowledge would come to them followed immediately by words, or if the “words are the knowledge?”, they said “Probably the words are the knowledge, [but it's] kind of a hard distinction to make”.
Not being able to readily use visual mental imagery didn’t seem to outwardly affect Taylor much. They described chemistry as involving a lot of mental visualisation and that they found it “Generally pretty easy, but it kind of depends I guess. […] [First-year chemistry] I found pretty easy and it got more harder when we got into harder visualisations in [second-year chemistry] […] which was a lot of the rings-type chemistry and it was like flipping it around and different things […] it was a lot of kind of trying to visualise exactly how to draw things.” They said that this was “where the molecular models came in quite handy, to figure out the shapes of it all”.
They made no indication that they felt their inability to easily use visual mental imagery was a hindrance in solving PSVT questions and said that being able to solve such questions was “very useful” for doing or learning chemistry – “[you] need some way to be able to do it, even if you’re not kind of visualising the whole thing. I think being able to rotate [objects], yeah, it's very useful. You could figure out ways to manage without it, but you’d be at a disadvantage.” Although Taylor exclusively used the verbal strategy to solve the PSVT questions, their response indicates that they still considered their process to involve mental rotation.
When asked if their good memory therefore translated to high academic performance, they said, “I think I have the capacity to be [a high achiever], but I think high achieving is the result of talent and effort. […] I would like to think I have the talent. I have a good capacity for memory. I don’t necessarily think that translates to problem solving and working things out.” They added, “I also have been recently lacking the effort because I did, with the memory, find that things were reasonably… how do I say this? Reasonably easy to retrieve and so study for me was reading notes, and if the effort wasn’t there, I didn’t have a lot of notes to read.”
They described that using visual mental imagery is “fairly easy” if it's “a single molecule for example, or something by itself, not really doing anything”, but that “once you start introducing complex interactions and tying in different things that it's doing, it becomes a bit more difficult for me. […] I think because then the focus is diverted. I feel like I can’t trust the accuracy of my visualisation. It's not to say that I can’t see it, […] I prefer to have some words there to back up what I’m seeing in my mind.”
They further elaborated how their vivid visual mental imagery did not translate into visuospatial ability: “It's just that me visualising things moving is not as easy as it is for me to just understand words and understand just the immediate picture, like something simple. So when it comes to expending mental effort, for like example, visualising rotation, moving these shapes and, example, moving myself through [a] path, is not as easy. […] My memory leans towards kind of just remembering things I’ve seen rather than problem solving.”
However, Bailey felt that visuospatial ability wasn’t an inherently fixed skill, as when asked if they felt their capacity to solve PSVT-style questions was reflected in their chemistry performance, they said, “Not necessarily. I think the ease at which these would come to you would reflect your ability to understand chemistry, but I think again this is something that with repeat exposure would get easier over time because it would be a matter of, you need learning to rotate and see things in 3D space, which I believe is a skill, I feel like you can pick that up. And skills have varying talents, you know, it's easier for some people. I think that the person who finds chemistry easier will find this easier.”
Discussion
The two linked studies reported here have explored the relationship between the strength of students’ mental imagery and their thinking as part of chemistry-related tasks which required complex visual and spatial reasoning.When asked to reflect on times they had visualised concepts in their mind, the interview participants referenced molecular structures, reaction mechanisms, visualising receptor–ligand bindings, enzyme reactions, and viral mechanisms. All of these self-reported concepts require understanding of chemistry-related spatial characteristics of systems, ranging from the atomic level to more complex systems such as cells and enzymes. This finding supported the rationale for our study into how mental imagery can affect visuospatial ability in chemistry.
To what extent does the strength of students’ mental imagery impact their success in thinking when dealing with representations in chemistry-related visuospatial tasks? (RQ 1)
Although the interview participants had different degrees of visual mental imagery strength, five out of seven reported that the difficulty of using visual mental imagery was different to the vividness of their visual mental imagery. In general, the participants described using visual mental imagery for simple concepts as straightforward, but that they experienced greater challenges as the task becomes more abstract or complex. For example, Bailey, who has strong visual mental imagery capabilities self-described as being akin to photographic memory, reported having much difficulty with mental manipulations in spatial tasks. Similarly, Taylor, who we earlier attributed to being aphantasic, described mental visualisation as “generally pretty easy” but more challenging with more complex concepts such as ring chemistry (more on this participant to follow in discussion of RQ 2). From this, we hypothesise that success in chemistry visuospatial thinking may be impacted by two separate dimensions: (1) vividness of visual imagery, and (2) ability to generate such imagery (regardless of its vividness).The SUVI task in the quantitative study was designed to assess whether students’ mental imagery influenced their speed and accuracy in selecting visual representations of functional groups. This task was adapted from Monzel and colleagues (2021) who found that individuals without aphantasia selected images of fruits faster than aphantasics, which they attributed to a priming effect of pre-emptively visualising the fruit. In contrast, our findings indicated no significant relationship between the strength of visual mental imagery (measured by the VOSI Visual-Object and Psi-Q Vision subscales) and the participants’ response speed in selecting either images of laboratory objects or chemistry functional groups. However, we observed that higher scores on the VOSI Visual-Spatial subcomponent were associated with slower but more accurate selections of functional group images. This finding provides some evidence for potential differences amongst students in the time they may require to complete tasks involving functional groups dependent on the strength of their mental imagery. Additionally, higher scores on the Psi-Q Bodily sensation subcomponent also correlated with increased accuracy in selecting functional group images. Together, these findings prompted further investigation of self-reported approaches to processing images by participants in the second study, specifically when we examined how they imagine functional groups. We explored whether their descriptions of how they imagined functional groups were at all related to their VOSI Visual-Spatial and Psi-Q Bodily sensation scores.
During interviews, considerable variability in how the participants imagined a carboxylic acid was observed. They described their mental construction process as involving between one to four steps. The time taken by the participants was positively but non-linearly correlated with the number of steps taken, meaning more steps took more time but each additional step contributed less time than the one prior. This aligns with findings in cognitive psychology on chunking, where individuals have been observed to reduce the processing time required to recall complex sequences by grouping elements into mental units (Wu et al., 2023). Contrary to our expectations as arising from our findings from the quantitative study, we found no association between the interview participants’ VOSI Visual-Spatial or Psi-Q Bodily sensation scores and either the time taken or the number of steps required to imagine the carboxylic acid. However, we did observe a significant negative relationship between their Psi-Q Vision scores and the time taken to imagine the carboxylic acid, indicating that participants with higher scores, suggestive of more vivid visual mental imagery, imagined this functional group faster. This result was unexpected and raised a question – if participants with vivid visual mental imagery imagine functional groups faster, why did they not select them faster as well in the first study? Regardless, this finding should be interpreted with caution, as six of the seven participants scored an 8 or higher out of 10 on the Psi-Q Vision subscale, resulting in limited variation within the data. While the small sample size prevents drawing firm conclusions, the finding represents an interesting area for future exploration.
Returning to some other findings from the quantitative diagnostic study, no relationships were found between participants’ vividness of visual mental imagery and their performance on PSVT:R_A, providing preliminary evidence suggesting that the ability to solve mental rotation questions may be independent of visual mental imagery strength. We observed that participants with higher scores on the VOSI Visual-Spatial, Psi-Q Sound, or Psi-Q Bodily sensation subscales were found to take longer times to solve questions involving two or more rotations of the object, but at no cost to accuracy. We hypothesised that more vivid mental imagery in these modalities would be associated with the use of more time-consuming strategies used to solve mental rotation questions. To explore this further, the interviews invited participants to describe their mental processes while solving several PSVT:R_A questions. We identified four distinct strategies used by the participants: Mental Rotation, Projection, Verbal, and Multiple. No association was found between mental imagery strength of the aforementioned mental imagery modalities and the strategy used by the participants. However, our analysis suggested a potential relationship between participants VOSI Visual-Object scores, which is intended to capture visual mental imagery strength, and their strategy choices. Specifically, our model indicated that individuals with higher VOSI Visual-Object scores tended to use Projection and Multiple strategies, whereas those with closer-to-mid-range scores predominantly used Mental Rotation. The only low-scoring interview participant exclusively employed the Verbal strategy and was the only participant to apply it. As with earlier findings, the conclusiveness of this result is limited by the small sample size – six of the seven interview participants scored 3.8/5 or higher on this subscale. Nevertheless, if we were to draw a tentative conclusion, it is that while visual mental imagery strength may influence strategy selection in mental rotation tasks, it may not impact task performance in terms of speed or accuracy, based on the collective findings from both the quantitative and qualitative studies. Further research is being undertaken to explore this, as well as to continue investigating the identified associations between spatial, sound and bodily sensation imagery.
Does aphantasia impact on students’ abilities to employ imagistic strategies in chemistry problem solving? (RQ 2)
During the interviews, one aim was to investigate whether aphantasic chemistry students can employ imagistic strategies during problem solving. We also sought to explore whether the concept of “visual-spatial imagery”, as discussed in the introduction, is applicable to individuals with aphantasia.In Case study 1, we explained that we believe Taylor to experience aphantasia in the sense of reduced voluntary visual mental imagery, rather than its absence entirely. Regardless of this distinction, their participation in interviews provided valuable insights. Throughout their interview, Taylor frequently used language that suggested the creation of visual mental images. However, they described that their images were neither vivid nor constructed well. Upon further inquiry, Taylor clarified that they experience mental “impressions” rather than visual mental images. These impressions involve imagining things without actually “seeing” them and are accompanied by verbal mental descriptions. Taylor exclusively used the Verbal strategy for solving the PSVT:R_A questions and was the only participant to use this strategy. The Verbal strategy involved mentally encoding verbal instructions and systemically analysing each rotation in a sequential manner. However, Taylor's description of their method suggested some elements of visual mental imagery, as suggested by a statement about object movement (“it’d have to move like that”, see Table 4). It remains unclear whether Taylor used visual mental imagery for these movements, or rather conceptualised them as “impressions”. This difference is important – while Taylor stored directional movement information via their strategy, the Mental Rotation strategy involves both storing and creating mental images of the objects in new positions. This distinction categorises the Verbal strategy as analytic and Mental Rotation as imagistic, according to Stieff (2007) and Stieff and colleagues (2010). Ultimately, Taylor could mentally manipulate spatial information – despite lacking readily useable visual mental imagery – and expressed appreciation for being able to do so “even if you’re not kind of visualising the whole thing”.
Stieff (2013) asserted that the time taken by individuals to solve mental rotation tasks correlates with angular disparity, but only when imagistic, not analytic, strategies are used. However, Pénzes (2023) and Kay and colleagues (2024) challenged this by showing that aphantasics’ reaction times on mental rotation tasks also increased with angular disparity. This suggests that either aphantasics can employ imagistic strategies, or that the assumption that a correlation between angular disparity and time inherently indicates the use of mental rotation requires revisiting. Our findings from the quantitative study, which uncovered no clear relationships between strength of visual imagery, angular disparity, and time taken, add to this ambiguity.
Furthermore, in the quantitative study, Taylor scored 1.2/5 on the VOSI Visual-Object subscale and 1.3/5 on the VOSI Visual-Spatial subscale. The only other participant to score below 2 on the Visual-Object subscale, with a score of 1.4/5, also scored 1.4/5 on the Visual-Spatial subscale. Hence, we are inclined to reject the notion that Blazhenkova's (2016) concept of “visual-spatial imagery” was intact for these participants. However, Taylor's descriptions of their mental processes, which included elements of abstract, possibly visual, mental imagery, introduces further uncertainty.
Ultimately, the question of whether individuals with aphantasia can utilise imagistic strategies remains unresolved, and is likely influenced by how one defines or interprets “imagistic strategies”. Further research is underway to explore this topic in greater depth.
Conclusions and implications for teaching
This paper has explored the relationship between students’ mental imagery strength and their thinking in relation to visuospatial problem solving in chemistry. Our findings suggest that the vividness of students’ visual mental imagery did not significantly impact their ability to spatially reason with visual representations in chemistry. We observed that a student with highly vivid visual mental imagery does not inherently have an advantage in learning chemistry. Similarly, a student with reduced voluntary visual mental imagery, who we considered as experiencing aphantasia, is not necessarily disadvantaged as they can use alternative strategies that do not rely on vivid visual mental imagery.We recommend against educators adopting a rigid position for analytic over imagistic teaching methods (or vice versa). Both our review of literature and results of our studies do not provide substantial evidence to support one being advantageous over another for all students. Additionally, we were unable to confirm whether aphantasic students could apply imagistic strategies. However, given the variability in individual experiences of aphantasia and the ambiguity surrounding definitions of imagistic strategies, we question the productivity of seeking a definitive answer. Instead, we advocate for educators to facilitate a learning environment that encourages students to develop and refine their own unique learning and problem-solving strategies under expert guidance. Such an inclusive approach supports and nurtures the diversity of unique cognitive profiles of all students.
Limitations and future studies
This study, while contributing novel quantitative and qualitative insights into the connection between mental imagery and learning in chemistry, has several limitations that require acknowledgement.• The small sample sizes in both studies likely do not capture the full diversity of imagistic ability existing in the broader population of chemistry students. We also opted not to collect detailed demographic data, such as gender and background. Together, these factors limit the generalisability of our findings. Additionally, there is the possibility that the statistically significant results reported may be due to Type I error. To overcome these limitations, our future studies aim to include a larger and more diverse sample size that can validate and extend our findings.
• This study's reliance on self-reported questionnaires (VOSI and Psi-Q) to assess mental imagery strength creates potential variability in our findings due to subjective interpretations of internal experiences by participants. Objective measures, such as neuroimaging techniques, could offer a more accurate assessment of mental imagery vividness. However, these approaches are challenging to access and costly to implement.
• The potential influence of unmeasured variables on the relationship between mental imagery strength and chemistry-related task performance. Factors such as prior knowledge in chemistry, representational competence, motivation, and other cognitive factors could significantly affect outcomes. While our research aimed to explore implicit associations between these factors and mental imagery strength, our findings ultimately show correlation and not causation, as such definitive conclusions cannot be made without further investigation. Therefore, future research should isolate and measure these variables, to better identify causality and extend our understanding of how mental imagery strength may influence chemistry learning.
• The digital tasks employed in the quantitative diagnostic surveys do not encompass the full range of visuospatial problem-solving scenarios that chemistry students may encounter. For example, our study used abstract 3D objects during the PSVT:R_A, which may not adequately parallel molecular representations in terms of visuospatial thinking. To address this, future research into mental imagery should explore the evaluation of students’ visuospatial abilities specifically in the context of molecular chemistry, offering a more focused understanding of representational competencies in our context.
• In this study we emphasised the potential necessity for educational approaches that suit students of differing mental imagery strengths. However, we did not offer any specific intervention strategies to address this. Future research could focus on the design, implementation and assessment of instructional methodologies that are created to cater to different levels of mental imagery ability. Such research would be invaluable in contribution to the development of more inclusive educational practices.
• The reliance on self-reported strategies by participant solving the PSVT:R_A questions, particularly by those who attempted to describe their process while solving the questions. Think-aloud reports may not fully capture the cognitive processes involved as they do not account for non-verbal and subconscious aspects of thought, and attempting to verbalise thoughts may modify participants’ natural thinking behaviour (Ericsson and Simon, 1998).
• The inclusion of only one participant with aphantasia in interviews limited our capacity to understand more thoroughly the cognitive processes of aphantasic chemistry students. Importantly, this participant was not classified as having ‘extreme’ aphantasia, in that their visual mental imagery was reduced and involuntary rather than absent entirely. The rarity of individuals with aphantasia poses a significant challenge for research, and is only worsened by low student research participation in general. Future research involving the recruitment of a larger sample of aphantasic chemistry students is essential.
• Finally, we operated under the assumption that vividness of mental imagery is constant amongst participants over time. However, the development and change of mental imagery skills over time, particularly after representational competencies are developed, remains unexplored. Longitudinal studies could provide insights into how and if mental imagery vividness changes over time, and the implications this has on learning.
We have recognised these limitations to emphasise the challenges faced in data collection to explore imagistic ability in a disciplinary context. However, the insights gained are important to inform chemistry education practice – research is underway to validate the findings in light of limitations.
Author contributions
Lauren Baade: conceptualisation, data curation, formal analysis, investigation, methodology, visualisation, writing – original draft & revisions. Effie Kartsonaki: conceptualisation, methodology, supervision. Hassan Khosravi: conceptualisation, methodology, supervision. Gwendolyn A. Lawrie: conceptualisation, methodology, supervision, writing – review and editing.Data availability
The data reported in this study have been collected applying ethical procedures approved by the human research ethical review committee at The University of Queensland. This data cannot be shared publicly to protect the privacy and confidentiality of the study's participants.Conflicts of interest
There are no conflicts of interest to declare.Acknowledgements
Lauren Baade was supported by the Australian Government Research Training Program (RTP) Tuition Fee Offset and Stipend. The authors wish to thank the students who participated in this study.References
- Ainsworth S., (2006), DeFT: a conceptual framework for considering learning with multiple representations, Learn. Instruct., 16(3), 183–198.
- Ainsworth S., (2014), The Multiple Representation Principle in Multimedia Learning, in: Mayer R. E. (ed.), The Cambridge Handbook of Multimedia Learning, 2nd edn, Cambridge: Cambridge University Press.
- Andrade J., May J., Deeprose C., Baugh, S. J. and Ganis, G., (2014), Assessing vividness of mental imagery: the Plymouth Sensory Imagery Questionnaire, Br. J. Psychol., 105(4), 547–563.
- Bainbridge W. A., Pounder Z., Eardley A. F. and Baker C. I., (2021), Quantifying Aphantasia through drawing: those without visual imagery show deficits in object but not spatial memory, Cortex, 135, 159–172.
- Bartlett K. A., (2023), The Politics of the Purdue Spatial Visualization Test of Rotations (PSVT: R) and its Use in Engineering Education, Eng. Stud., 1(1), 1–22.
- Bartoń K., (2023), MuMIn: Multi-Model Inference. 1.47.5 ed.
- Bates D., Mächler M., Bolker B. and Walker S., (2015), Fitting linear mixed-effects models using lme4, J. Stat. Softw., 67(1), 1–48.
- Beran M. J., James B. T., French K., Haseltine E. L. and Kleider-Offutt H. M., (2023), Assessing aphantasia prevalence and the relation of self-reported imagery abilities and memory task performance, Conscious. Cogn., 113, 103548.
- Bergmann J., Genç E., Kohler A., Singer W. and Pearson J., (2016), Smaller primary visual cortex is associated with stronger, but less precise mental imagery, Cereb. Cortex, 26(9), 3838–3850.
- Blajenkova O., Kozhevnikov M. and Motes M. A., (2006), Object-spatial imagery: a new self-report imagery questionnaire, Appl. Cogn. Psychol., 20(2), 239–263.
- Blazhenkova O., (2016), Vividness of Object and Spatial Imagery, Percept. Mot. Skills, 122(2), 490–508.
- Blazhenkova O. and Pechenkova E., (2019), The two eyes of the blind mind: object vs. spatial aphantasia? Russ. J. Cogn. Sci., 6(4), 51–65.
- Blomkvist A. and Marks D. F., (2023), Defining and ‘diagnosing’aphantasia: Condition or individual difference? Cortex, 169, 220–234.
- Boone Jr H. N. and Boone D. A., (2012), Analyzing likert data, J. Ext., 50(2), 48.
- Bumgardner A. L., Yuan K. and Chiu A. V., (2021), I cannot picture it in my mind: acquired aphantasia after autologous stem cell transplantation for multiple myeloma, Oxf. Med. Case Rep., 2021(5), omab019.
- Carlisle D., Tyson J. and Nieswandt M., (2015), Fostering spatial skill acquisition by general chemistry students, Chem. Ed. Res. Prac., 16(3), 478–517.
- Carter C. S., Larussa M. A. and Bodner G. M., (1987), A study of two measures of spatial ability as predictors of success in different levels of general chemistry. J. Res. Sci. Teach., 24(7), 645–657.
- Chase W. G. and Simon H. A., (1973), Perception in chess. Cogn. Psychol., 4(1), 55–81.
- Cole M., Wilhelm J., Vaught B. M.-M., Fish C. and Fish H., (2020), The relationship between spatial ability and the conservation of matter in middle school, Educ. Sci., 11(1), 4.
- Cooper M. M. and Stowe R. L., (2018), Chemistry education research—From personal empiricism to evidence, theory, and informed practice, Chem. Rev., 118(12), 6053–6087.
- Cui X., Jeter C. B., Yang D., Montague P. R. and Eagleman D. M., (2007), Vividness of mental imagery: individual variability can be measured objectively, Vision Res., 47(4), 474–478.
- Dance C., Ipser A. and Simner J., (2022), The prevalence of aphantasia (imagery weakness) in the general population, Conscious. Cogn., 97, 103243.
- Dawes A. J., Keogh R., Andrillon T. and Pearson J., (2020), A cognitive profile of multi-sensory imagery, memory and dreaming in aphantasia, Sci. Rep., 10(1), 1–10.
- Dawes A. J., Keogh R. and Pearson J., (2024), Multisensory subtypes of aphantasia: Mental imagery as supramodal perception in reverse, Neurosci. Res., 201, 50–59.
- De Boeck P. and Jeon M., (2019), An overview of models for response times and processes in cognitive tests, Front. Psychol., 10, 422756.
- Eme P.-E. and Marquer J., (1999), Individual strategies in a spatial task and how they relate to aptitudes, Eur. J. Psychol. Educ., 14, 89–108.
- Ericsson K. A. and Simon H. A., (1998), How to study thinking in everyday life: contrasting think-aloud protocols with descriptions and explanations of thinking, Mind Cult. Act., 5(3), 178–186.
- Fox J. and Weisberg S., (2018), An R companion to applied regression, Sage Publications.
- Fulford J., Milton F., Salas D., Smith A., Simler A., Winlove C. and Zeman A., (2018), The neural correlates of visual imagery vividness – An fMRI study and literature review, Cortex, 105, 26–40.
- Gałecki A., Burzykowski T., Gałecki A. and Burzykowski T., (2013), Linear mixed-effects model, Springer.
- Galton F., (1883), Mental imagery. Inquiries into human faculty and its development. New York, NY, US: MacMillan Co.
- Graulich N., Langner A., Vo K. and Yuriev E., (2021), Scaffolding metacognition and resource activation during problem solving: a continuum perspective, Probl. Probl. Solving Chem. Educ.: Analysing Data, Look. Patterns Make. Deductions, 1, 38–67.
- Greco S., Ishizaka A., Tasiou M. and Torrisi G., (2019), On the methodological framework of composite indices: a review of the issues of weighting, aggregation, and robustness, Soc. Indic. Res., 141(61), 61–94.
- Gulyás E., Gombos F., Sütöri S., Lovas A., Ziman G. and Kovács I., (2022), Visual imagery vividness declines across the lifespan, Cortex, 154, 365–374.
- Hall K. W., Kouroupis A., Bezerianos A., Szafir D. A. and Collins C., (2021), Professional differences: a comparative study of visualization task performance and spatial ability across disciplines, IEEE Trans. Vis. Comput. Graph., 28(1), 654–664.
- Hegarty M., Stieff M. and Dixon B. L., (2013), Cognitive change in mental models with experience in the domain of organic chemistry, J. Cogn. Psychol., 25(2), 220–228.
- Hensen C. and Barbera J., (2019), Assessing affective differences between a virtual general chemistry experiment and a similar hands-on experiment, J. Chem. Educ., 96(10), 2097–2108.
- Hornbuckle S. F., Gobin L. and Thurman S. N., (2014), Spatial Reasoning: Improvement of Imagery and Abilities in Sophomore Organic Chemistry. Perspective to Enhance Student Learning, Contemp. Issues Educ. Res., 7(1), 45–50.
- Johnstone A. H., (1982), Macro and microchemistry, Chem. Brit., 18(6), 409–410.
- Kampen J. and Swyngedouw M., (2000), The ordinal controversy revisited, Qual. Quant., 34(1), 87–102.
- Kay L., Keogh R. and Pearson J., (2024). Slower but more accurate mental rotation performance in aphantasia linked to differences in cognitive strategies, Conscious. Cogn., 121, 103694.
- Kell H. J. and Lubinski D., (2013), Spatial ability: a neglected talent in educational and occupational settings, Roeper Rev., 35(4), 219–230.
- Keogh R. and Pearson J., (2018), The blind mind: No sensory visual imagery in aphantasia, Cortex, 105, 53–60.
- Keogh R. and Pearson J., (2021). Attention driven phantom vision: measuring the sensory strength of attentional templates and their relation to visual mental imagery and aphantasia, Philos. Trans. R. Soc. B, 376(1817), 20190688.
- Keogh R., Wicken M. and Pearson J., (2021), Visual working memory in aphantasia: retained accuracy and capacity with a different strategy, Cortex, 143, 237–253.
- Kiernan N. A., Manches A. and Seery M. K., (2021), The role of visuospatial thinking in students’ predictions of molecular geometry, Chem. Educ. Res. Pract., 22(3), 626–639.
- Kiernan N. A., Manches A. and Seery M. K., (2024), Resources for reasoning of chemistry concepts: multimodal molecular geometry, Chem. Educ. Res. Pract., 25(2), 524–543.
- Kozma R. and Russell J., (1997), Multimedia and understanding: Expert and novice responses to different representations of chemical phenomena. J. Res. Sci. Teach., 34(9), 949–968.
- Kozma R. and Russell J., (2005), Students becoming chemists: developing representational competence, Visualization in Science Education, Springer, p. 123.
- Lalla M., (2017), Fundamental characteristics and statistical analysis of ordinal variables: a review, Qual. Quant., 51, 435–458.
- Leech N. L. and Onwuegbuzie A. J., (2009), A typology of mixed methods research designs, Qual. Quant, 43, 265–275.
- Lenth R. V., (2023), emmeans: Estimated Marginal Means, aka Least-Squares Means. 1.8.9 ed.
- Leutner D., Leopold C. and Sumfleth E., (2009), Cognitive load and science text comprehension: effects of drawing and mentally imagining text content, Comput. Hum. Behav., 25(2), 284–289.
- Liu J. and Bartolomeo P., (2023), Probing the unimaginable: The impact of aphantasia on distinct domains of visual mental imagery and visual perception, Cortex, 166, 338–347.
- Maeda Y. and Yoon S. Y., (2013), A meta-analysis on gender differences in mental rotation ability measured by the Purdue spatial visualization tests: Visualization of rotations (PSVT: R), Educ. Psychol. Rev., 25(1), 69–94.
- Marks D. F., (1973), Visual imagery differences in the recall of pictures, Br. J. Psychol., 64(1), 17–24.
- Miller G. A., (1956), The magical number seven, plus or minus two: Some limits on our capacity for processing information, Psychol. Rev., 63(2), 81.
- Milton F., Fulford J., Dance C., Gaddum J., Heuerman-Williamson B., Jones K., Knight K. F., Mackisack M., Winlove C. and Zeman A., (2021), Behavioral and neural signatures of visual imagery vividness extremes: aphantasia versus hyperphantasia, Cereb. Cortex Commun., 2(2), tgab035.
- Mitrenga K. J., Alderson-Day B., May L., Moffatt J., Moseley P. and Fernyhough C., (2019), Reading characters in voices: Ratings of personality characteristics from voices predict proneness to auditory verbal hallucinations, PLoS One, 14(8), e0221127.
- Monzel M., Keidel K. and Reuter M., (2021), Imagine, and you will find–Lack of attentional guidance through visual imagery in aphantasics, Atten. Percept. Psychophys., 83, 2486–2497.
- Nanay B., (2018), Multimodal mental imagery, Cortex, 105, 125–134.
- Pearson J. and Kosslyn S. M., (2015), The heterogeneity of mental representation: Ending the imagery debate, Proc. Natl. Acad. Sci. U. S. A., 112(33), 10089–10092.
- Pénzes D., (2023), Interpreting Mental Rotation Performance in Self-Described Aphantasia through Cognitive Penetrability, Independent thesis, Umeå University.
- Pinheiro J. C. and Bates D. M., (2000). Linear mixed-effects models: basic concepts and examples, Mixed-effects models S S-Plus, 3–56.
- Pounder Z., Jacob J., Evans S., Loveday C., Eardley A. F. and Silvanto J., (2022). Only minimal differences between individuals with congenital aphantasia and those with typical imagery on neuropsychological tasks that involve imagery, Cortex, 148, 180–192.
- Pribyl J. R. and Bodner G. M., (1987), Spatial ability and its role in organic chemistry: a study of four organic courses, J. Res. Sci. Teach., 24(3), 229–240.
- Rabionet S. E., (2011), How I learned to design and conduct semi-structured interviews: an ongoing and continuous journey, Qual. Rep., 16(2), 563–566.
- Rademaker R. L. and Pearson J., (2012), Training visual imagery: Improvements of metacognition, but not imagery strength, Front. Psychol., 3, 224.
- Rapp D. N., (2005), Mental models: theoretical issues for visualizations in science education, Visualization in science education, pp. 43–60.
- Ratcliff R., (1993), Methods for dealing with reaction time outliers, Psychol. Bull., 114(3), 510.
- Rau M. A., (2016), A framework for educational technologies that support representational competencies, IEEE Trans. Learn. Technol., 10(3), 290–305.
- Rau M. A., (2017), Conditions for the effectiveness of multiple visual representations in enhancing STEM learning, Educ. Psychol. Rev., 29(4), 717–761.
- Rau M. A., Aleven V. and Rummel N., (2017), Making connections among multiple graphical representations of fractions: sense-making competencies enhance perceptual fluency, but not vice versa, Instr. Sci., 45(3), 331–357.
- R Core Team, (2024), R: A Language and Environment for Statistical Computing, Vienna, Australia: R Foundation for Statistical Computing.
- Robin X., Turck N., Hainard A., Tiberti N., Lisacek F., Sanchez J.-C. and Müller M., (2011), pROC: an open-source package for R and S+ to analyze and compare ROC curves, BMC Bioinform., 12(1), 1–8.
- Saariluoma P. and Kalakoski V., (1997), Skilled imagery and long-term working memory, Am. J. Psy., 110(2), 177–201.
- Saldana J., (2011), Fundamentals of qualitative research, Oxford University Press.
- Schack T., Essig K., Frank C. and Koester D., (2014), Mental representation and motor imagery training, Front. Hum. Neurosci., 8, 328.
- Schnotz W., (2005), An integrated model of text and picture comprehension, The Cambridge Handbook of Multimedia Learning.
- Schnotz W., (2014), Integrated Model of Text and Picture Comprehension, The Cambridge Handbook of Multimedia Learning.
- Seufert T., (2003), Supporting coherence formation in learning from multiple representations, Learn. Instr., 13(2), 227–237.
- Southam D. C. and Lewis J. E., (2013), Supporting alternative strategies for learning chemical applications of group theory, J. Chem. Educ., 90(11), 1425–1432.
- Stieff M., (2007), Mental rotation and diagrammatic reasoning in science, Learn. Instr., 17(2), 219–234.
- Stieff M., (2011), When is a molecule three dimensional? A task-specific role for imagistic reasoning in advanced chemistry, Sci. Educ., 95(2), 310–336.
- Stieff M., (2013), Sex differences in the mental rotation of chemistry representations, J. Chem. Educ., 90(2), 165–170.
- Stieff M., Dixon B. L., Ryu M., Kumi B. C. and Hegarty M., (2014), Strategy training eliminates sex differences in spatial problem solving in a STEM domain. J. Educ. Psychol., 106(2), 390.
- Stieff M., Hegarty M. and Dixon B., (2010), Alternative strategies for spatial reasoning with diagrams, Proc. Int. Conf. Theory Appl. Diagrams, Springer, pp. 115–127.
- Stieff M., Origenes A., Desutter D., Lira M., Banevicius L., Tabang D. and Cabel G., (2018), Operational constraints on the mental rotation of STEM representations, J. Educ. Psychol., 110(8), 1160.
- Stieff M. and Raje S., (2008), Expertise and spatial reasoning in advanced scientific problem solving, in Kirschner P. A., van Merriënboer J. J. G. and de Jong T. (ed.), Proceedings of the Eighth International Conference of the Learning Sciences, Mahwah, NJ: Erlbaum, vol. 2, pp. 366–373.
- Stieff M. and Raje S., (2010), Expert algorithmic and imagistic problem solving strategies in advanced chemistry, Spatial Cogn. Comput., 10(1), 53–81.
- Stieff M., Ryu M., Dixon B. and Hegarty M., (2012), The role of spatial ability and strategy preference for spatial problem solving in organic chemistry, J. Chem. Educ., 89(7), 854–859.
- Stieff M. and Uttal D., (2015), How much can spatial training improve STEM achievement? Educ. Psychol. Rev., 27, 607–615.
- Stieff M., Werner S., Desutter D., Franconeri S. and Hegarty M., (2020), Visual chunking as a strategy for spatial thinking in STEM, Cogn. Res. Princ. Implic., 5(115), 1–15.
- Stoet G., (2010), PsyToolkit: a software package for programming psychological experiments using Linux, Behav. Res. Methods, 42(4), 1096–1104.
- Stoet G., (2017), PsyToolkit: a novel web-based method for running online questionnaires and reaction-time experiments, Teach. Psychol., 44(1), 24–31.
- Tomasino B., Pellitteri G., Bax F., Marini A., Surcinelli A., Gigli G. L. and Valente M., (2022), Multisensory mental representation in covid-19 patients and the possibility of long-lasting gustatory and olfactory dysfunction in the CNS, Sci. Rep., 12(1), 7340.
- U. S. Bureau of Labor Statistics, (2023), Employment by major occupational group [Online]. Available: https://www.bls.gov/emp/tables/emp-by-major-occupational-group.htm [Accessed 2023-12-04].
- Uttal D. H. and Doherty K. O., (2008), Comprehending and learning from ‘visualizations’: A developmental perspective, Visualization: Theory and Practice in Sci. Educ., Dordrecht: Springer Netherlands, pp. 53–72.
- Villanueva R. A. M. and Chen Z. J., (2019), ggplot2: elegant graphics for data analysis, Taylor & Francis.
- Vishnumolakala V. R., Southam D. C., Treagust D. F., Mocerino M. and Qureshi S., (2017), Students’ attitudes, self-efficacy and experiences in a modified process-oriented guided inquiry learning undergraduate chemistry classroom, Chem. Educ. Res. Pract., 18(2), 340–352.
- Vlacholia M., Vosniadou S., Roussos P., Salta K., Kazi S., Sigalas M. and Tzougraki C., (2017), Changes in visual/spatial and analytic strategy use in organic chemistry with the development of expertise, Chem. Educ. Res. Pract., 18(4), 763–773.
- Vo K., Sarkar M., White P. J. and Yuriev E., (2022), Problem solving in chemistry supported by metacognitive scaffolding: teaching associates’ perspectives and practices, Chem. Educ. Res. Pract., 23(2), 436–451.
- Waters A. J. and Gobet F., (2008), Mental imagery and chunks: empirical and computational findings, Mem. Cog., 36, 505–517.
- Whelan R., (2008), Effective analysis of reaction time data, Psychol. Rec., 58, 475–482.
- Willis S. G., Wheatley G. H. and Mitchell O. R., (1979), Cerebral processing of spatial and verbal-analytic tasks: an EEG study, Neuropsychologia, 17(5), 473–484.
- Wu S., Éltető N., Dasgupta I. and Schulz E., (2023), Chunking as a rational solution to the speed–accuracy trade-off in a serial reaction time task, Sci. Rep., 13(1), 7680.
- Yang E.-M., Andre T., Greenbowe T. J. and Tibell L., (2003), Spatial ability and the impact of visualization/animation on learning electrochemistry, Int. J. Sci. Educ., 25(3), 329–349.
- Yoon S. Y., (2011), Psychometric properties of the revised purdue spatial visualization tests: visualization of rotations (The Revised PSVT: R), Purdue University.
- Zeman A., (2024), Aphantasia and hyperphantasia: exploring imagery vividness extremes, Trends Cognitive Sci., 28(5), 467–480.
- Zeman A. Z. J., Della Sala S., Torrens L. A., Gountouna V.-E., McGonigle D. J. and Logie R. H., (2010), Loss of imagery phenomenology with intact visuo-spatial task performance: a case of ‘blind imagination’, Neuropsychologia, 48(1), 145–155.
- Zeman A. Z., Dewar M. and Della Sala S., (2015), Lives without imagery-Congenital aphantasia, Cortex, 73, 378–380.
- Zeman A. Z., Dewar M. T. and Della Sala S., (2016), Reflections on aphantasia, Cortex, 74, 336–337.
- Zeman A., Milton F., Della Sala S., Dewar M., Frayling T., Gaddum J., Hattersley A., Heuerman-Williamson B., Jones K. and Mackisack M., (2020), Phantasia – the psychological significance of lifelong visual imagery vividness extremes, Cortex, 130, 426–440.
Footnote |
| † Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d4rp00234b |
| This journal is © The Royal Society of Chemistry 2025 |