Allocentric memory impaired and egocentric memory intact as assessed by virtual reality in recent-onset schizophrenia

Present evidence suggests that schizophrenia is associated with explicit memory deficits, whereas implicit memory seems to be largely preserved. Virtual reality studies on declarative allocentric memory in schizophrenia are rare, and studies on implicit egocentric memory in schizophrenia are lacking. However, virtual realities have a major advantage for the assessment of spatial navigation and memory formation, as computer-simulated first-person environments can simulate navigation in a large-scale space. Twenty-five subjects with recent-onset schizophrenia were compared with 25 healthy matched control subjects on two virtual reality tasks affording the navigation and learning of a virtual park (allocentric memory) and a virtual maze (egocentric memory). Compared with control subjects, schizophrenia subjects were significantly impaired in learning the virtual park. However, schizophrenia subjects were as able as control subjects to learn the virtual maze. Stronger disorganized symptoms of schizophrenia subjects were significantly related to more errors on the virtual maze. It is concluded that egocentric spatial learning adds to the many other implicit cognitive skills being largely preserved in schizophrenia. Possibly, the more global neural network supporting egocentric spatial learning is less affected than the declarative hippocampal memory system in early stages of schizophrenia and may offer opportunities for compensation in the presence of focal deficits.     

Egocentric memory impaired and allocentric memory intact as assessed by virtual reality in subjects with unilateral parietal cortex lesions

Present evidence suggests that medial temporal cortices subserve allocentric representation and memory, whereas egocentric representation and memory mainly depends on inferior and superior parietal cortices. Virtual reality environments have a major advantage for the assessment of spatial navigation and memory formation, as computer-simulated first-person environments can simulate navigation in a large-scale space. However, virtual reality studies on allocentric memory in subjects with cortical lesions are rare, and studies on egocentric memory are lacking. Twenty-four subjects with unilateral parietal cortex lesions due to infarction or intracerebral haemorrhage (14 left-sided, 10 right-sided) were compared with 36 healthy matched control subjects on two virtual reality tasks affording to learn a virtual park (allocentric memory) and a virtual maze (egocentric memory). Subjects further received a comprehensive clinical and neuropsychological investigation, and MRI lesion assessment using T₁, T₂ and FLAIR sequences as well as 3D MRI volumetry at the time of the assessment. Results indicate that left- and right-sided lesioned subjects did not differ on task performance. Compared with control subjects, subjects with parietal cortex lesions were strongly impaired learning the virtual maze. On the other hand, performance of subjects with parietal cortex lesions on the virtual park was entirely normal. Volumes of the right-sided precuneus of lesioned subjects were significantly related to performance on the virtual maze, indicating better performance of subjects with larger volumes. It is concluded that parietal cortices support egocentric navigation and imagination during spatial learning in large-scale environments.     

Egocentric and allocentric memory as assessed by virtual reality in individuals with amnestic mild cognitive impairment

Present evidence suggests that medial temporal cortices subserve allocentric representation and memory, whereas egocentric representation and memory also depends on parietal association cortices and the striatum. Virtual reality environments have a major advantage for the assessment of spatial navigation and memory formation, as computer-simulated first-person environments can simulate navigation in a large-scale space. Twenty-nine patients with amnestic MCI (aMCI) were compared with 29 healthy matched controls on two virtual reality tasks affording to learn a virtual park (allocentric memory) and a virtual maze (egocentric memory). Participants further received a neuropsychological investigation and MRI volumetry at the time of the assessment. Results indicate that aMCI patients had significantly reduced size of the hippocampus bilaterally and the right-sided precuneus and inferior parietal cortex. aMCI patients were severely impaired learning the virtual park and the virtual maze. Smaller volumes of the right-sided precuneus were related to worse performance on the virtual maze. Participants with striatal lacunar lesions committed more errors than participants without such lesions on the virtual maze but not on the virtual park. aMCI patients later converting to dementia (n = 15) had significantly smaller hippocampal size when compared with non-converters (n = 14). However, both groups did not differ on virtual reality task performance. Our study clearly demonstrates the feasibility of virtual reality technology to study spatial memory deficits of persons with aMCI. Future studies should try to design spatial virtual reality tasks being specific enough to predict conversion from MCI to dementia and conversion from normal to MCI.     

l-DOPA changes ventral striatum recruitment during motor sequence learning in Parkinson's disease

Recent research indicates that longstanding temporal lobe epilepsy (TLE) is associated with extratemporal, i.e. parietal cortex damage. We investigated egocentric and allocentric memory by use of first-person large-scale virtual reality environments in patients with TLE. We expected that TLE patients with parietal cortex damage were impaired in the egocentric memory task. Twenty-two TLE patients with hippocampal sclerosis (HS) and 22 TLE patients without HS were compared with 42 healthy matched controls on two virtual reality tasks affording to learn a virtual park (allocentric memory) and a virtual maze (egocentric memory). Participants further received a neuropsychological investigation and MRI volumetry at the time of the assessment. When compared with controls, TLE patients with HS had significantly reduced size of the ipsilateral and contralateral somatosensory cortex (postcentral gyrus). When compared with controls or TLE patients without HS, TLE patients with HS were severely impaired learning the virtual maze. Considering all participants, smaller volumes of the left-sided postcentral gyrus were related to worse performance on the virtual maze. It is concluded that the paradigm of egocentric navigation and learning in first-person large-scale virtual environments may be a suitable tool to indicate significant extratemporal damage in individuals with TLE.     

Posterior parahippocampal gyrus lesions in the human impair egocentric learning in a virtual environment

Functional imaging studies have shown that the posterior parahippocampal gyrus (PHG) is involved in allocentric (world-centered) object and scene recognition. However, the putative role of the posterior PHG in egocentric (body-centered) spatial memory has received only limited systematic investigation. Thirty-one subjects with pharmacoresistant medial temporal lobe epilepsy (TLE) and temporal lobe removal were compared with 19 matched healthy control subjects on a virtual reality task affording the navigation in a virtual maze (egocentric memory). Lesions of the hippocampus and PHG of TLE subjects were determined by three-dimensional magnetic resonance imaging volumetric assessment. The results indicate that TLE subjects with right-sided posterior PHG lesions were impaired on virtual maze acquisition when compared with controls and TLE subjects with anterior PHG lesions. Larger posterior PHG lesions were significantly related to stronger impairments in virtual maze performance. Our results point to a role of the right-sided posterior PHG for the representation and storage of egocentric information. Moreover, access to both allocentric and egocentric streams of spatial information may enable the posterior PHG to construct a global and comprehensive representation of spatial environments.     

Performance on a virtual reality spatial memory navigation task in depressed patients

OBJECTIVE: Findings on spatial memory in depression have been inconsistent. A navigation task based on virtual reality may provide a more sensitive and consistent measure of the hippocampal-related spatial memory deficits associated with depression. METHOD: Performance on a novel virtual reality navigation task and a traditional measure of spatial memory was assessed in 30 depressed patients (unipolar and bipolar) and 19 normal comparison subjects. RESULTS: Depressed patients performed significantly worse than comparison subjects on the virtual reality task, as assessed by the number of locations found in the virtual town. Between-group differences were not detected on the traditional measure. The navigation task showed high test-retest reliability. CONCLUSIONS: Depressed patients performed worse than healthy subjects on a novel spatial memory task. Virtual reality navigation may provide a consistent, sensitive measure of cognitive deficits in patients with affective disorders, representing a mechanism to study a putative endophenotype for hippocampal function.     

A comparison of egocentric and allocentric spatial memory in medial temporal lobe and Korsakoff amnesics

Two patients with medial temporal lobe damage, seven Korsakoff amnesics and fourteen healthy control subjects were tested on three conditions of a spatial memory test ('short delay', 'allocentric' and 'egocentric'). The task required subjects to recall the position of a single spot of light presented on a board after various delays. The 'short delay' condition tested memory over very short, unfilled intervals. The other two conditions used longer, filled delays. The allocentric condition required subjects to move to a different place around the board before recalling the position of the light. In the egocentric condition stimuli were presented in darkness, which eliminated allocentric cues. The Korsakoff amnesics were impaired at all delays of the short delay tasks, suggesting poor encoding. On the allocentric and egocentric tasks the Korsakoff amnesics showed a comparable impairment in the two conditions, which worsened with delay. This accelerated forgetting suggested that the Korsakoff amnesics also had impaired memory for allocentric and egocentric information. The patients with medial temporal lobe damage were unimpaired in the 'short delay' condition suggesting intact encoding and short-term memory of spatial information. However, they were impaired in the allocentric condition and showed accelerated loss of allocentric spatial information. In the egocentric condition, while the performance of one patient was impaired, the performance of the other was as good as controls. This result suggests that, in contrast to allocentric spatial memory, which is sensitive to medial temporal lobe damage, an intact medial temporal lobe need not be necessary for successful performance on an egocentric spatial memory task.     

As the world turns: short-term human spatial memory in egocentric and allocentric coordinates

We aimed to determine whether human subjects' reliance on different sources of spatial information encoded in different frames of reference (i.e., egocentric versus allocentric) affects their performance, decision time and memory capacity in a short-term spatial memory task performed in the real world. Subjects were asked to play the Memory game (a.k.a. the Concentration game) without an opponent, in four different conditions that controlled for the subjects' reliance on egocentric and/or allocentric frames of reference for the elaboration of a spatial representation of the image locations enabling maximal efficiency. We report experimental data from young adult men and women, and describe a mathematical model to estimate human short-term spatial memory capacity. We found that short-term spatial memory capacity was greatest when an egocentric spatial frame of reference enabled subjects to encode and remember the image locations. However, when egocentric information was not reliable, short-term spatial memory capacity was greater and decision time shorter when an allocentric representation of the image locations with respect to distant objects in the surrounding environment was available, as compared to when only a spatial representation encoding the relationships between the individual images, independent of the surrounding environment, was available. Our findings thus further demonstrate that changes in viewpoint produced by the movement of images placed in front of a stationary subject is not equivalent to the movement of the subject around stationary images. We discuss possible limitations of classical neuropsychological and virtual reality experiments of spatial memory, which typically restrict the sensory information normally available to human subjects in the real world.     

Time course of allocentric decay, egocentric decay, and allocentric-to-egocentric conversion in memory-guided reach

Allocentric cues can be used to encode locations in visuospatial memory, but it is not known how and when these representations are converted into egocentric commands for behaviour. Here, we tested the influence of different memory intervals on reach performance toward targets defined in either egocentric or allocentric coordinates, and then compared this to performance in a task where subjects were implicitly free to choose when to convert from allocentric to egocentric representations. Reach and eye positions were measured using Optotrak and Eyelink Systems, respectively, in fourteen subjects. Our results confirm that egocentric representations degrade over a delay of several seconds, whereas allocentric representations remained relatively stable over the same time scale. Moreover, when subjects were free to choose, they converted allocentric representations into egocentric representations as soon as possible, despite the apparent cost in reach precision in our experimental paradigm. This suggests that humans convert allocentric representations into egocentric commands at the first opportunity, perhaps to optimize motor noise and movement timing in real-world conditions.     

Sequential egocentric strategy is acquired as early as allocentric strategy: Parallel acquisition of these two navigation strategies

At least two main cognitive strategies can be used to solve a complex navigation task: the allocentric or map‐based strategy and the sequential egocentric or route‐based strategy. The sequential egocentric strategy differs from a succession of independent simple egocentric responses as it requires a sequential ordering of events, possibly sharing functional similarity with episodic memory in this regard. To question the possible simultaneous encoding of sequential egocentric and allocentric strategies, we developed a paradigm in which these two strategies are spontaneously used or imposed. Our results evidenced that sequential egocentric strategy can be spontaneously acquired at the onset of the training as well as allocentric strategy. Allocentric and sequential egocentric strategies could be used together within a trial, and bidirectional shifts (between trials) were spontaneously performed during the training period by 30% of the participants. Regardless of the strategy used spontaneously during the training, all participants could execute immediate shifts to the opposite non previously used strategy when this strategy was imposed. Altogether, our findings suggest that subjects acquire different types of spatial knowledge in parallel, namely knowledge permitting allocentric navigation as well as knowledge permitting sequential egocentric navigation. © 2009 Wiley‐Liss, Inc.     

Real world referencing and schizophrenia: Are we experiencing the same reality?

Psychotic symptoms in schizophrenia patients encompass the difficulty to distinguish between the respective points of view of self and others. The capacity to adopt and switch between different perspectives is, however, fundamental for ego- and allocentric spatial referencing. We tested whether schizophrenia patients are able to adopt and maintain a non-egocentric point of view in a complex visual environment. Twenty-four chronic schizophrenic outpatients (11 females) and 25 controls matched for age, gender, years of education and handedness were recruited from a population-based sample. In a virtual environment, participants had to make a decision as to which of two trash cans was closest to themselves (viewer-centered, egocentric), to a ball (object-centered, unstable allocentric), or to a palace (landmark-centered, stable allocentric). Main outcome measures were reaction time, error rate, learning rate and local task switch cost. While egocentric reaction time was preserved, patients showed an increased reaction time in both allocentric referencing conditions (stable and unstable) and an overall increased error rate. Switch cost was diminished in patients when changing from the egocentric to the landmark-centered condition and elevated when changing from the landmark-centered to the egocentric condition. The results imply that schizophrenia patients’ adoption of an egocentric perspective is preserved. However, adopting an allocentric point of view and switching between egocentric and landmark-centered perspectives are impaired. Perturbations in non-egocentric referencing and transferring efficiently between different referential systems might contribute to altered personal and social world comprehension in schizophrenia.     

Disrupted allocentric but preserved egocentric spatial learning in transgenic mice with impaired glucocorticoid receptor function

Spatial and non-spatial learning of mice with an incorporated antisense RNA complementary to a fragment of cDNA coding for the glucocorticoid receptor (GR) were evaluated in allocentric and egocentric radial maze and water maze tasks, and in spontaneous object recognition and sensorimotor learning paradigms. Mice with impaired GR function did not acquire two maze paradigms based on allocentric spatial navigation, radial maze non-matching to position and water maze spatial discrimination learning. Comparison of performance in spaced and massed trials indicated that this may be due to a general inability to store information into allocentric reference memory or in retrieval processes. However, both groups of animals learned the rules of an egocentric radial maze task at similar rates and there was no difference in their ability to recognise objects once animals had equal opportunity to explore the sample objects. Sensorimotor performance was impaired in transgenic animals, but it is suggested that this is due to non-specific factors rather than to disrupted sensorimotor learning per se. These results are consistent with a disruption of hippocampal function. Histological examination of the hippocampus revealed no obvious structural abnormalities in transgenic animals. Therefore, the data suggest that functional underactivity of GRs at the level of the hippocampus induces a deficit in allocentric navigation while sparing egocentric navigation and object recognition.     

Recognition memory for hand positions and spatial locations in patients with Huntington's disease: differential visuospatial memory impairment?

Allocentric and egocentric memory was investigated in patients with Huntington's disease (HD) and matched controls. Patients with HD and age- and education-matched healthy normal controls (NC) were administered two visuospatial recognition memory tasks, one assessing memory for hand positions (egocentric) and the other assessing memory for spatial locations (allocentric). HD patients showed normal primacy and recency effects, but their overall performance was impaired relative to controls on both tasks. Correlation analyses indicated that HD patients' performance on the Hand Position Memory task, but not the Spatial Location Memory task, was associated with global cognitive status (Mattis Dementia Rating Scale) and disease severity (Shoulson and Fahn Rating Scale), and HD patients' performances on the two tasks were not associated. Results provide preliminary support for the role of the caudate nucleus in both allocentric and egocentric spatial memory.     

Differential working memory impairment in bipolar disorder and schizophrenia: effects of lifetime history of psychosis

BACKGROUND: Although bipolar disorder and schizophrenia have long been viewed as distinct illnesses, there is growing evidence that these two complex diseases share some common genes, which may manifest as overlapping neuropsychological impairments. Although working memory dysfunction has been proposed to be central to the pathophysiology of schizophrenia, it has received less attention in studies of bipolar disorder. METHOD: We applied measures of working memory to patients with schizophrenia (n = 15), patients with schizoaffective disorder (n = 15), patients with psychotic (n = 11) and non-psychotic (n = 15) bipolar disorder, and demographically matched healthy subjects (n = 32), in order to determine the extent to which these groups show common or unique impairments. RESULTS: While patients with bipolar disorder (with and without psychotic features) and those with schizophrenia/schizoaffective disorder were impaired on backward digit span, only patients with a lifetime history of psychotic features, regardless of diagnosis, were impaired on spatial delayed response task. CONCLUSIONS: Backward digit span performance is comparable in bipolar disorder and schizophrenia, and may be an appropriate endophenotypic marker that cuts across diagnostic categories. In contrast, spatial working memory performance clearly distinguishes non-psychotic bipolar disorder patients from patients with functional psychosis.     

Prefrontal function and activation in bipolar disorder and schizophrenia

OBJECTIVE: Distinctive patterns of speech and language abnormalities are associated with bipolar disorder and schizophrenia. It is, however, unclear whether the associated patterns of neural activation are diagnosis specific. The authors sought to determine whether there are differences in language-associated prefrontal activation that discriminate bipolar disorder and schizophrenia. METHOD: Forty-two outpatients with bipolar I disorder, 27 outpatients with schizophrenia, and 37 healthy comparison subjects were recruited. Differences in blood oxygen level-dependent activity were evaluated using the Hayling Sentence Completion Test and analyzed in Statistical Parametric Mapping (SPM) 2. Differences in activation were estimated from a sentence completion versus rest contrast and from a contrast of decreasing sentence constraint. Regional activations were related to clinical variables and performance on a set shifting task and evaluated for their ability to differentiate among the three groups. RESULTS: Patients with bipolar disorder showed differences in insula and dorsal prefrontal cortex activation, which differentiated them from patients with schizophrenia. Patients with bipolar disorder recruited the orbitofrontal cortex and ventral striatum to a greater extent relative to healthy comparison subjects on the parametric contrast of increasing difficulty. The gradient of ventral striatal and prefrontal activation was significantly associated with reversal errors in bipolar disorder patients. CONCLUSIONS: Brain activations during the Hayling task differentiated patients with bipolar disorder from comparison subjects and patients with schizophrenia. Patients with bipolar disorder showed abnormalities in frontostriatal systems associated with performance on a set shifting task. This finding suggests that bipolar disorder patients engaged emotional brain areas more than comparison subjects while performing the Hayling task.     

Cholinergic receptor blockade in prefrontal cortex and lesions of the nucleus basalis: implications for allocentric and egocentric spatial memory in rats

In this study we have examined the involvement of the prefrontal cortex (PFC) along with the Nucleus basalis magnocellularis (NBM) in two types of spatial navigation tasks. We evaluated the effects of excitotoxic (ibotenate-induced) lesions of the NBM in an allocentric and an egocentric task in the Morris water maze, using sham operations for a comparison. In both cases we also assessed the effects of local cholinergic receptor blockade in the PFC by infusing the muscarinic receptor antagonist scopolamine (4 or 20 microg). Anatomically, the results obtained showed that this lesion produced a profound loss of acetylcholinesterase (AChE) positive cells in the NBM, and a loss of AChE positive fibres in most of the neocortex, but hardly in the medial PFC. Behaviourally, such lesions led to a severe impairment in the allocentric task. Intraprefrontal infusions of scopolamine led to a short-lasting impairment in task performance when the high dose was used. In the second experiment, using the same surgical manipulations, we examined the performance in the egocentric task. Like in the allocentric task animals with NBM lesions were also impaired, but with continued training they acquired a level of performance similar to the sham-operated ones. This time, infusions of scopolamine in the medial PFC led to a severe disruption of performance in both groups of animals. We conclude that acetylcholine in the medial PFC is important for egocentric but not allocentric spatial memory, whereas the NBM is involved in the learning of both tasks, be it to a different degree.     

Glucose-induced increase in memory performance in patients with schizophrenia.

Previous investigations have found that increasing circulating glucose availability can increase memory performance in rodents, healthy humans, and individuals with dementia of the Alzheimer's type. In this study, patients with schizophrenia, healthy control subjects, and controls with bipolar affective disorder were tested using double-blind treatment with either 50 g anhydrous dextrose plus 4 mg sodium saccharin (for "taste") or 23.7 mg saccharin alone, followed by cognitive testing on a complex battery. At this glucose dose, verbal memory performance on a paragraph recall task was increased during the glucose condition relative to the saccharin condition in the patients with schizophrenia; this effect was not detected in either the psychiatric or normal controls. The results provide preliminary support for the hypothesis that memory performance can be improved in patients with schizophrenia by increasing circulating glucose availability and suggest the importance of further evaluation of therapeutic manipulations of glucose availability.     

Using virtual reality to characterize episodic memory profiles in amnestic mild cognitive impairment and Alzheimer's disease: influence of active and passive encoding

Most neuropsychological assessments of episodic memory bear little similarity to the events that patients actually experience as memories in daily life. The first aim of this study was to use a virtual environment to characterize episodic memory profiles in an ecological fashion, which includes memory for central and perceptual details, spatiotemporal contextual elements, and binding. This study included subjects from three different populations: healthy older adults, patients with amnestic mild cognitive impairment (aMCI) and patients with early to moderate Alzheimer's disease (AD). Second, we sought to determine whether environmental factors that can affect encoding (active vs. passive exploration) influence memory performance in pathological aging. Third, we benchmarked the results of our virtual reality episodic memory test against a classical memory test and a subjective daily memory complaint scale. Here, the participants were successively immersed in two virtual environments; the first, as the driver of a virtual car (active exploration) and the second, as the passenger of that car (passive exploration). Subjects were instructed to encode all elements of the environment as well as the associated spatiotemporal contexts. Following each immersion, we assessed the patient's recall and recognition of central information (i.e., the elements of the environment), contextual information (i.e., temporal, egocentric and allocentric spatial information) and lastly, the quality of binding. We found that the AD patients' performances were inferior to that of the aMCI and even more to that of the healthy aged groups, in line with the progression of hippocampal atrophy reported in the literature. Spatial allocentric memory assessments were found to be particularly useful for distinguishing aMCI patients from healthy older adults. Active exploration yielded enhanced recall of central and allocentric spatial information, as well as binding in all groups. This led aMCI patients to achieve better performance scores on immediate temporal memory tasks. Finally, the patients' daily memory complaints were more highly correlated with the performances on the virtual test than with their performances on the classical memory test. Taken together, these results highlight specific cognitive differences found between these three populations that may provide additional insight into the early diagnosis and rehabilitation of pathological aging. In particular, neuropsychological studies would benefit to use virtual tests and a multi-component approach to assess episodic memory, and encourage active encoding of information in patients suffering from mild or severe age-related memory impairment. The beneficial effect of active encoding on episodic memory in aMCI and early to moderate AD is discussed in the context of relatively preserved frontal and motor brain functions implicated in self-referential effects and procedural abilities.     

Spatial working memory and planning ability: contrasts between schizophrenia and bipolar I disorder

Working memory may be conceptualized as a multi-component system involving the active maintenance and manipulation of stored information in the service of planning/guiding behaviour. Impaired spatial working memory is a robust finding in schizophrenia patients which has been related to an impairment in frontostriatal connectivity. The purpose of this study was to examine the specificity of this impairment by comparing the mnemonic and executive aspects of working memory performance in schizophrenia and bipolar disorder with psychotic features, focusing particularly on the functional dynamics between task components. Twenty-four patients with schizophrenia, 14 patients with bipolar I disorder (manic phase) and 33 healthy control subjects were assessed using the Cambridge Neuropsychological Test Automated Battery (CANTAB): including the spatial working memory (between search errors and strategy scores) spatial span (storage capacity) and spatial planning (Stockings of Cambridge: accuracy and latency) tasks. Both patient groups were impaired on the spatial span task, which requires the maintenance and retrieval of stored information. In contrast, only schizophrenia patients showed a significant deficit in between search errors, which requires both maintenance and manipulation of information in working memory. That is, they exhibited both a mnemonic and an executive dysfunction. Spatial span was particularly important to accurate planning ability in bipolar patients. In contrast, in patients with schizophrenia poor spatial working memory was a significant predictor of planning impairments, consistent with failures in goal selection, evaluation and/or execution. Furthermore, initial planning time was positively correlated with the latency to complete a planning sequence. This pattern of slow cognitive processing in schizophrenia patients only, resembled that reported previously in patients with basal ganglia disorders. These findings are discussed in terms of a possible common disturbance in fronto-parietal circuitry in the two disorders together with a specific disturbance of fronto-striatal circuitry in schizophrenia, that is not present in bipolar disorder.     

Physiology of schizophrenia, bipolar disorder, and schizoaffective disorder

OBJECTIVE: Endophenotypes have been proposed to identify the genetic and biological substrates of complex disorders. Three physiological inhibitory endophenotypes of large effect size in schizophrenia include suppression of P50 auditory evoked responses, inhibition of leading (small anticipatory) saccades during smooth pursuit eye movements, and cancellation of reflexive saccades in the antisaccade eye movement task. The aim of this study was to determine if the pattern of endophenotype abnormalities within individuals with schizophrenia differed from that within individuals with bipolar disorder. A second aim was to determine whether subjects with schizoaffective disorder, bipolar type, were neurophysiologically more similar to subjects with schizophrenia or subjects with bipolar disorder. METHOD: Endophenotypes were recorded for subjects diagnosed with schizophrenia (N=29), bipolar disorder (DSM-IV-TR) (N=40), and schizoaffective disorder, bipolar type (N=18). Data from normal comparison subjects were used to establish normal performance. RESULTS: Logistic regression determined that P50 ratio and frequency of leading saccades identified subjects with schizophrenia and bipolar disorder with a sensitivity of 95% and a specificity of 83%. The schizoaffective disorder group was split, with six subjects physiologically classified as schizophrenia-like and 12 subjects as bipolar-like. Those classified as schizophrenia-like were significantly younger at illness onset and had higher symptom ratings. CONCLUSION: A composite endophenotype of P50 ratio and frequency of leading saccades is consistent with the current clinical nosology of schizophrenia and bipolar disorder and parses patients with schizoaffective disorder, bipolar type, into two subgroups.