Edited by: Zbigniew R. Struzik, The University of Tokyo, Japan
Reviewed by: Joel Pearson, The University of New South Wales, Australia; Ahmad Abu-Akel, University of Birmingham, UK
*Correspondence: Taylor L. Benson, Department of Psychology, Vanderbilt University, 111 21st Ave S, Nashville, TN 37240, USA e-mail:
This article was submitted to the journal Frontiers in Human Neuroscience.
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
Biographical and historical accounts suggest a link between scientific creativity and schizophrenia. Longitudinal studies of gifted children indicate that visuospatial imagery plays a pivotal role in exceptional achievements in science and mathematics. We asked whether visuospatial imagery is enhanced in individuals with schizophrenia (SZ). We compared SZ and matched healthy controls (HC) on five visuospatial tasks tapping parietal and frontoparietal functions. Two aspects of visuospatial transformation, spatial location and mental imagery manipulation were examined with Paper Folding Test (PFT) and jigsaw puzzle task (JPT), respectively. Visuospatial intelligence was assessed with Ravens Progressive Matrices, which is associated with frontoparietal network activity. Hemispatial inattention implicating parietal function was assessed with line bisection (LB) task. Mediated by prefrontal cortex, spatial delayed response task (DRT) was used to index working memory maintenance, which was impaired in SZ compared to HC. In contrast, SZ showed intact visuospatial intelligence and transformation of location. Further, SZ performed significantly better than HC on JPT indicating enhanced mental imagery manipulation. Spatial working memory (SWM) maintenance and mental imagery manipulation were strongly associated in HC but dissociated in SZ. Thus, we observed enhanced mental imagery manipulation in SZ but the dissociation of mental imagery from working memory suggests a disrupted frontoparietal network. Finally, while HC showed the expected leftward pseudoneglect, SZ showed increased rightward LB bias implicating left hemispatial inattention and impaired right parietal control of spatial attention. The current results chart a unique profile of impaired, spared and enhanced parietal-mediated visuospatial functions implicating parietal abnormalities as a biobehavioral marker for SZ. We discuss these results in relation to creative cognition.
Two large-scale, longitudinal studies of gifted children in the past half-century have sought to answer the question of what lies at the core of exceptional abilities. Project TALENT (
Mental images are the building blocks of human consciousness that comprise one's internal representation of the world. Historical and biographical accounts have often suggested the pivotal role of mental imagery in exceptional creativity. For instance, Albert Einstein rarely thought in words, preferring instead to rely on images (Wertheimer,
Einstein's reliance on imagery is believed to have played a major role in his thought processes and discoveries. Indeed, Einstein's special theory of relativity had its roots in a thought experiment that he had been visualizing since his adolescence. As a boy, he wanted to know what a beam of light would look like if he could ride alongside it. So he imagined himself traveling at the speed of light with a mirror held in front of him. In this imaginary scenario, he could never see his reflection in the mirror since the light and the mirror were traveling at the same velocity and the mirror was always held a little ahead of the beam. Such use of mental images while working on new theories mark Einstein's thinking style.
Much has been speculated about the nature and source of Einstein's genius. Initial analyses of his brain structures had not yielded anything remarkable but the recent discovery of lost slides of Einstein's brain and subsequent re-analyses of cortical region by Falk et al. (
A second unique feature of Einstein's brain was the asymmetry and the size of his parietal cortex, which plays a central role in visuospatial functions, especially mental imagery, mathematical ability and multisensory processing. Einstein's right superior parietal lobule (SPL) was strikingly different from his left SPL. The third unusual aspect of his brain involves the sensory motor cortices. An area of the motor cortex that processes information from the face, tongue, and larynx was expanded into a large rectangular patch that apparently has not been observed in other brains (see Falk et al.,
An important footnote to the case of Einstein is his genetic liability for schizophrenia as a first-degree relative. Genetic or psychometric high-risk for schizophrenia has been linked to elevated creativity in epidemiological (Kyaga et al.,
These autobiographical anecdotes are confirmed by research (Forisha,
A creative idea must be both novel and appropriate (Sternberg and O'Hara,
In this paper, we examine the link between mental imagery (MI), creativity and psychopathology to move beyond these fascinating anecdotal case studies and show why we need neuroscience to understand the origins of human creativity. Previous research has emphasized the relationship between mental imagery and at least two fundamental features of schizophrenia (SZ), a severely debilitating psychotic disorder characterized by widespread cognitive impairments that negatively impact functional outcome (Green et al.,
Working memory is an active, limited-capacity, short-term memory system, which temporarily maintains information, and provides an interface between perception, long-term memory and action (Baddeley,
Patients with schizophrenia consistently demonstrate impaired performance on behavioral tasks requiring the utilization of SWM ability (Park and Holzman,
In contrast to the cognitive deficits observed in the schizophrenia spectrum, recent research indicates that SZ patients show spared abilities in certain cognitive domains, and sometimes even enhanced performance when compared to healthy controls (HC). Understanding these rare enhancements in schizophrenia is important since they may lead to a more complete understanding of the complex cognitive profile of patients with schizophrenia, rather than just describing a generalized deficit in cognitive performance. Such enhancements include self-reported vividness of mental imagery (e.g., Sack et al.,
Mental images are similar to actual percepts in terms of their functions in interference tasks, image scanning times, and brain activation patterns (Kosslyn,
Since both mental imagery and working memory have been hypothesized to be markers for schizophrenia (i.e., enhanced mental imagery as a trait marker and impaired working memory as an endophenotype candidate), it is interesting that these two cognitive processes show opposite patterns of results (i.e., enhanced mental imagery and impaired working memory). The puzzle becomes even more complex in light of the realization that mental imagery and working memory are both cognitive processes that rely on the maintenance and manipulation of internal representations (Kosslyn,
Despite this intriguing dissociation, however, only a handful of studies have empirically investigated the dissociation between mental imagery and working memory in SZ patients. The few that have attempted such a feat in the past have mostly focused on parsing out maintenance vs. manipulation components of the SWM deficit. Thakkar and Park (
Similarly, Matthews et al. (
One potential benefit that might result from enhanced mental imagery includes increased creative achievement, which has been found to be linked with the schizophrenia-spectrum (Abraham et al.,
One possibility is the presence of pockets of enhanced creativity in probands of the disorder (O'Reilly et al.,
The goal of the current investigation was to help reconcile the deficits and enhancements that have been previously reported in the cognitive profile of SZ patients. We sought to accomplish this broad goal by specifically examining visuospatial mental imagery manipulation ability and its relationship with SWM in SZ patients compared with demographically matched HC. Given that HC show strong positive correlation between mental imagery and SWM, we wanted to further investigate the dissociation between these two cognitive processes in SZ. Finding evidence for this hypothesized MI/SWM dissociation in the current SZ sample would further implicate frontoparietal connectivity abnormalities as a neural marker that may be important for both SZ and creative cognition.
Eighteen medicated outpatients with schizophrenia (SZ) were recruited from a private psychiatric facility in Nashville, TN. Diagnoses were confirmed using the structured clinical interview for DSM-IV (SCID-IV; First et al.,
Age | 0.6761 | |||
Sex | 5 females, 13 males | 8 females, 10 males | χ2(1) = 1.08 | 0.2979 |
Handedness | 15 right, 3 non-right | 17 right, 1 non-right | χ2(1) = 1.172 | 0.2791 |
Edinburgh laterality | 0.1841 | |||
Premorbid IQ (NART) | 0.8416 | |||
Years of education | 0.0174 |
|||
SPQ (total) | <0.0001 |
|||
PDI (total) | 0.0002 |
|||
PDI (distress) | <0.0001 |
|||
PDI (preoccupation) | 0.0005 |
|||
PDI (conviction) | 0.0006 |
|||
PDI (paranoia) | 0.0012 |
|||
PDI (thought disturbance) | 0.0029 |
|||
PDI (catastrophic thought broadcast) | 0.0010 |
|||
PDI (ideation of reference) | 0.0079 |
|||
BPRS (total) | ||||
SAPS (total) | ||||
SANS (total) | ||||
CPZ-EQ (mg/kg/day) | ||||
Duration of illness (years) | ||||
Number of hospitalizations |
Current symptom ratings for SZ patients were assessed on the day of testing using the Brief Psychiatric Rating Scale (BPRS; Overall and Gorham,
SWM ability was assessed with a spatial delayed response task (DRT) (see Park et al.,
Visuospatial intelligence was assessed with Raven's Standard Progressive Matrices (RPM) (Raven et al.,
Visuospatial transformation ability was assessed with the Paper Folding Test (PFT) (ETS,
Mental imagery manipulation was assessed with a jigsaw puzzle task (JPT) similar to Richardson and Vecchi's (
Prior to seeing the puzzle pieces, each participant was asked to visualize a certain object before they were presented with a scrambled, fragmented puzzle of the object they were previously instructed to visualize. The participants were given an answer sheet with the exact same grid as the puzzle and were instructed to fill in the numbers of the corresponding puzzle pieces to determine the correct orientation of the puzzle (see Figure
Spatial neglect was assessed with a standard paper-and-pencil version of the line bisection (LB) task (Schenkenberg et al.,
Previous research utilizing the LB task has found that spatial neglect is associated with lesions in the right parietal lobule (e.g., Mort et al.,
A MANCOVA was used to compare group mean performances on the battery of visuospatial tasks, using education as a covariate. The multivariate result was significant for group, Pillai's Trace = 0.560,
Spatial DRT accuracy (% correct) | 4.381 | 0.044 | 0.120 | 0.528 | SZ < HC | ||
Raven's matrices accuracy (% rank) | 0.311 | 0.581 | 0.009 | 0.084 | SZ = HC | ||
Raven's matrices time (minutes) | 0.129 | 0.721 | 0.004 | 0.064 | SZ = HC | ||
Paper folding test accuracy (% correct) | 0.015 | 0.903 | <0.001 | 0.052 | SZ = HC | ||
Jigsaw puzzle task accuracy (% correct) | 9.307 | 0.004 | 0.220 | 0.842 | SZ > HC | ||
Jigsaw puzzle task time (seconds) | 10.27 | 0.003 | 0.237 | 0.875 | SZ < HC | ||
Jigsaw puzzle task imagery agreement | 2.371 | 0.133 | 0.067 | 0.321 | SZ = HC | ||
Jigsaw puzzle task familiarity ratings | 1.736 | 0.197 | 0.050 | 0.249 | SZ = HC | ||
Jigsaw puzzle task complexity ratings | 0.837 | 0.367 | 0.025 | 0.144 | SZ = HC | ||
Line bisection task magnitude of bias | 9.678 | 0.004 | 0.227 | 0.855 | SZ > HC | ||
Line bisection task # right deviations | 6.493 | 0.016 | 0.164 | 0.696 | SZ > HC | ||
Line bisection task # left deviations | 6.927 | 0.013 | 0.173 | 0.724 | SZ < HC |
SWM accuracy | |||||||||||
Raven's % rank | SZ: 0.57 |
||||||||||
Raven's time | SZ: 0.18 HC: −0.42 | SZ: 0.01 HC: −0.10 | |||||||||
PFT accuracy | SZ: 0.08 HC: 0.59 |
SZ: 0.64 |
SZ: −0.29 HC:−0.29 | ||||||||
JPT accuracy | SZ: 0.16 HC: 0.64 |
SZ: 0.167 HC: 0.77 |
SZ: −0.23 HC: −0.32 | SZ: 0.01 HC: 0.74 |
|||||||
JPT time | SZ: 0.03 HC: −0.74 |
SZ: −0.16 HC: −0.72 |
SZ: 0.49 |
SZ: −0.48 |
SZ: −0.46 HC: −0.9 |
||||||
LB bias | SZ: 0.03 HC: −0.05 | SZ: −0.13 HC: 0.15 | SZ: −0.13 HC: 0.35 | SZ: 0.09 HC: −0.12 | SZ: −0.09 HC: −0.06 | SZ: −0.167 HC: 0.08 | |||||
PDI total | SZ: 0.04 HC: −0.65 |
SZ: −0.36 HC: −0.23 | SZ: 0.008 HC: 0.43 | SZ: −0.06 HC: −0.49 |
SZ: −0.29 HC: −0.37 | SZ: 0.14 HC: 0.46 | SZ: 0.47 |
||||
BPRS total | SZ: 0.11 HC: n/a | SZ: 0.04 HC: n/a | SZ: 0.13 HC: n/a | SZ: 0.11 HC: n/a | SZ: −0.15 HC: n/a | SZ: 0.11 HC: n/a | SZ: 0.11 HC: n/a | SZ: 0.48 |
|||
SAPS total | SZ: 0.15 HC: n/a | SZ: 0.12 HC: n/a | SZ: 0.07 HC: n/a | SZ: 0.23 HC: n/a | SZ: −0.18 HC: n/a | SZ: 0.16 HC: n/a | SZ: 0.03 HC: n/a | SZ: 0.61 |
SZ:0.76 |
||
SANS total | SZ: −0.18 HC: n/a | SZ: 0.12 HC: n/a | SZ: −0.01 HC: n/a | SZ: 0.28 HC: n/a | SZ: 0.11 HC: n/a | SZ: −0.19 HC: n/a | SZ: −0.09 HC: n/a | SZ: −0.05 HC: n/a | SZ: 0.52 |
SZ: 0.05 HC: n/a |
Since both groups performed poorly on the PFT (see Table
A chi-square contingency table was created and a subsequent Fisher's Exact Test was conducted to compare previous puzzle experience between SZ and HC. There was a trend-level difference [χ2(1) = 3.7,
Among SZ, LB bias was positively correlated with total scores of the PDI (
Mental imagery is intimately linked to memory (Baddeley and Andrade,
Previous research suggests that SZ patients consistently demonstrate an impaired ability to maintain mental representations, despite a superior ability to generate, inspect, and manipulate mental imagery (Thakkar and Park,
SZ and HC showed equal performance on the PFT, indicating spared visuospatial transformation ability in SZ. Furthermore, SZ and HC showed equal accuracy and completion time for RPM. The fact that we found no significant group differences between SZ and HC on RPM indicates that the two groups were matched for visuospatial intelligence. This is an important point since the two groups were recruited with the intent to be matched for premorbid intelligence without accounting for current intelligence in our SZ patient sample. Thus, equal performances on Raven's indicates that these two groups were matched on relevant levels of intelligence, suggesting that any group differences that we found on the other three administered tasks were due to variability inherent between the two groups not confounded by differences in IQ.
Consistent with previous literature, we found increased LB bias magnitude (in general) and greater sum of right deviations in particular, in patients with SZ compared to HC. This finding provides further evidence implicating the right parietal cortex as an important site for neural anomalies in the schizophrenia-spectrum. Consistent with recent reports implicating right LB bias as a potential marker for an endophenotype candidate for the schizophrenia-spectrum, we also found behavioral evidence for possible right parietal abnormalities in patients with SZ. However, LB bias in both SZ and HC, particularly magnitude of right deviations, was strongly related to delusional ideation in the current sample. Indeed, previous neuroimaging research has demonstrated a relationship between passivity delusions and right parietal abnormalities (Maruff et al.,
SZ performed significantly better than HC on the JPT. Specifically, SZ successfully completed the jigsaw puzzles in less time, and with fewer errors, than demographically matched HC. Further, SZ's JPT performance was unrelated to symptom severity, medication dosage, or illness duration. Consistent with Sack et al. (
SZ did not differ from HC in their post-task puzzle ratings. The fact that we found no significant group differences between SZ and HC on the complexity and familiarity ratings indicates that there were no prior biases that may have placed SZ at a greater advantage for solving the puzzles faster and more accurately than controls (other than superior imagery manipulation ability). The fact that there was no group difference on the image agreement ratings could mean one of two things: either the patients were able to use a strategy other than superior imagery generation to outperform HC on these tasks, or that the JPT used in the current study is a more sensitive measure of mental imagery than self-report methods. It is unclear from the results of this study which explanation is more plausible. However, at least one group of studies also found similar results suggesting superior mental imagery abilities via behavioral methodology, but not through self-reported methods (e.g., Matthews et al.,
Mental imagery and SWM abilities were strongly correlated in HC. In contrast, these two functions are dissociated in SZ, which may indicate abnormal connectivity of the frontoparietal network. Previous research has demonstrated that the parietal and frontoparietal networks may be important for creative cognition (e.g., Gaser and Schlaug,
The current results also suggest a potential cognitive remediation strategy for SZ. In other words, superior mental imagery manipulation ability in SZ could be leveraged to support working memory during cognitive training. This hypothesis is supported by preliminary evidence in the current study when one considers the discrepancy between presence vs. absence of group differences between the PFT and the JPT. Specifically, participants were repeatedly instructed to utilize mental imagery when solving the jigsaw puzzles, which is contrasted with the PFT in which participants were only instructed to use imagery during the initial instructions. Future research should explicitly test the hypothesis that a failure of connection between mental imagery and working memory regions could underlie some of the cognitive deficits and psychotic symptoms observed in SZ patients.
One potential limitation of the current study was that the two groups were not matched for years of education or previous puzzle experience. In detail, HC had significantly more years of education than SZ, in addition to higher levels of previous experience with jigsaw puzzles, particularly in the most recent 5 years prior to testing. Specifically, HC reported slightly more recent experience with jigsaw puzzles as a result of playing similar games with their children. However, despite greater experience with jigsaw puzzles and higher education level, the HC group did not perform better than SZ on the mental imagery tasks. Future research should attempt to match these demographic characteristics to help ensure that any group differences are not likely due to extraneous confounds. Still, as these discrepancies favor HC over SZ in the present study (i.e., HC had more education), the superior performance of SZ cannot be explained by these variables and could even be interpreted as a potential strengthening factor in light of the current results. In other words, even with less puzzle experience and a significant deficit in educational training, SZ still managed to outperform HC on the JPT, presumably by utilizing their superior visuospatial imagery manipulation abilities.
In sum, the results from the current investigation utilizing behavioral spatial cognition tasks suggest that in addition to showing evidence of parietal abnormalities, patients with SZ demonstrate impaired maintenance component of SWM, intact visuospatial intelligence and transformation, and enhanced visuospatial manipulation. SWM and mental imagery were positively correlated in HC but were essentially unrelated in SZ. This pattern of results is consistent with previous literature (Matthews et al.,
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
This research was supported by MH073028 and HD15052. We would like to thank Heath Nichols, Katy Thakkar, Lindsey Gilling McIntosh, and Joel Peterman for subject recruitment and symptoms ratings, and Mike Geoghegan for his helpful feedback on the manuscript.