Altered perception of apparent motion in schizophrenia spectrum disorder

Available online at www.sciencedirect.com

Psychiatry Research 159 (2008) 290 – 299 www.elsevier.com/locate/psychres

Altered perception of apparent motion in schizophrenia spectrum disorder Wolfgang Tschacher a,⁎, Priscilla Dubouloz b , Rahel Meier b , Uli Junghan a a

University Hospital of Psychiatry, University of Bern, Laupenstrasse 49, CH-3010 Bern, Switzerland b University Hospital (Inselspital), University of Bern, Murtenstrasse 35, 3010 Bern, Switzerland Received 24 May 2006; received in revised form 20 November 2006; accepted 7 April 2007

Abstract Apparent motion (AM), the Gestalt perception of motion in the absence of physical motion, was used to study perceptual organization and neurocognitive binding in schizophrenia. Associations between AM perception and psychopathology as well as meaningful subgroups were sought. Circular and stroboscopic AM stimuli were presented to 68 schizophrenia spectrum patients and healthy participants. Psychopathology was measured using the Positive and Negative Syndrome Scale (PANSS). Psychopathology was related to AM perception differentially: Positive and disorganization symptoms were linked to reduced gestalt stability; negative symptoms, excitement and depression had opposite regression weights. Dimensions of psychopathology thus have opposing effects on gestalt perception. It was generally found that AM perception was closely associated with psychopathology. No difference existed between patients and controls, but two latent classes were found. Class A members who had low levels of AM stability made up the majority of inpatients and control subjects; such participants were generally young and male, with short reaction times. Class B typically contained outpatients and some control subjects; participants in class B were older and showed longer reaction times. Hence AM perceptual dysfunctions are not specific for schizophrenia, yet AM may be a promising stage marker. © 2007 Elsevier Ireland Ltd. All rights reserved. Keywords: Diagnostic marker; Gestalt perception; Neurocognition; Perceptual organization; Psychopathology; Schizophrenia

1. Introduction Apparent motion (AM) is the perception of motion when physical motion of the stimuli presented is absent. A familiar example of AM is the stroboscopic effect used in cinematography. Although only a rapid succession of stills is displayed on the screen, smooth, uninterrupted motion (a “movie”) is seen. Gestalt theory posited AM as a demonstration of the holistic aspects of perception (“phi phenomenon” in Wertheimer, 1912). ⁎ Corresponding author. Tel.: +41 31 3876111; fax: +41 31 3829020. E-mail address: [email protected] (W. Tschacher).

The simplest paradigm that evokes AM consists in presenting a stimulus, e.g. a black disk, alternately in positions A and B of the visual field. Over a wide range of interstimulus intervals and frequencies of flashing, the viewer perceives the stimulus wandering back and forth between A and B. Clearly, the perceptual quality of movement is actively added to the physical stimuli by the viewer's information-processing system. AM represents the constructive properties of perception, together with further phenomena of perceptual organization, such as perceptual grouping and figure-ground discrimination. AM phenomena are thus not only prerequisite to being able to watch movies, but are also ecologically

0165-1781/$ - see front matter © 2007 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.psychres.2007.04.005

W. Tschacher et al. / Psychiatry Research 159 (2008) 290–299

significant for the functioning of visual perception in general. The visual scanning of scenes is commonly performed in successions of single fixations and intermittent fast saccadic eye movements. Nevertheless, the optical scene is subjectively seen as stationary during scanning, although the eyes may move repeatedly. The reason for perceived stationarity in spite of massive retinal translations is the reafference system (von Holst and Mittelstaedt, 1950), a feedback loop by which the brain compensates for bodily motion and eye movement. Viewing “real” locomotion, with eyes moving in a saccadic fashion, and viewing “apparent” motion in a laboratory paradigm can hence be considered similar from the standpoint of neurocognitive processing. Consequently, AM perception is a valid model for perception of motion in natural settings, and any alteration in the former should have an impact on perception in general. An exact functional description of anatomical regions of the brain involved in the perception of AM is still lacking. The raw input to visual processing consists of an unstructured array of features, which is actively segmented and grouped to reach a visual object representation (Mitroff and Scholl, 2005; Tschacher et al., 2006). Available research suggests that AM perception is located in brain sites higher than the primary visual cortex; a complex cortical network of motion-sensitive areas is driven by bottom-up and topdown neural processes (Goebel et al., 1998). Magnetic resonance imaging studies showed that areas in the middle temporal and the middle superior temporal regions in the dorsal processing stream of the visual cortex (MT (V5), MST) as well as prefrontal areas are involved. The investigation of visual perception via AM can be disentangled from other aspects of the processing of motion stimuli. It was found, for instance, that schizophrenic patients have impaired eye movement (Chen et al., 1999). AM, however, is perceived with eyes fixated. Hence confounding of AM perception with ocular motor problems is unlikely, especially when paradigms are used in which fixation of eyes during the trials is instructed. 1.1. Gestalt perception in schizophrenia spectrum disorders Gestalt psychological applications in psychiatry have a long tradition, notably in endeavours to better understand schizophrenia. Pioneers in this approach were Matussek (1952) and Conrad (1958), who considered the core phenomenology of schizophrenia

291

as arising from a dysfunctional processing of the gestalt patterns afforded by stimuli. In subsequent neuropsychological research, this idea was operationalized by various tasks requiring perceptual organization. In the following, a brief reference is made to a few of the many studies. Schizophrenia research using AM paradigms has been performed by Saucer and Deabler (1956), who found lowered organizational ability in schizophrenia. They investigated how AM perception changed and gave way to a flickering impression when flashing frequency was increased; these thresholds of loss of coherent AM discriminated between organic and schizophrenic patients, who both had low, deficient thresholds, and normal controls. Saucer (1958) replicated this finding for undifferentiated schizophrenia patients, yet found no deficiency in paranoid schizophrenia. Chambers and Wilson (1968) reported differences between schizophrenia patients and controls in an AM discrimination task. Place and Gilmore (1980) implemented a line-counting task to assess the influence of perceptual grouping on performance. Schizophrenia patients were not affected by the level of organization of lines, suggesting that they benefited to a lower degree from gestalt patterns than controls. Correspondingly, Silverstein et al. (2000) found that the ability of schizophrenia patients to detect contours (realized by regularly arranged Gabor patches) was reduced when the contours were embedded in “noisy” backgrounds (additional patches distributed randomly). A majority of studies showed impairments of perceptual organization in schizophrenia (Brand et al., 2005; Kéri et al., 2005; Silverstein and Uhlhaas, 2004; Uhlhaas and Silverstein, 2003, 2005), indicating that many persons with schizophrenia appear to have deficient gestalt perception. This specific kind of cognitive dysfunction was linked with neurobiological processes, especially with the functioning of cortical NMDA receptors. Phillips and Silverstein (2003) argued that an under-activity of NMDA glutamate receptor channels may be consequential for impaired cognitive coordination in psychotic states. In addition to differences between schizophrenia patients and control groups, several studies have reported significant correlations between perceptual organization and dimensions of psychopathology, such as, for instance, the disorganization syndrome (Silverstein et al., 2000), positive symptoms (Goodarzi et al., 2000) and negative symptoms (Doniger et al., 2001). In the Chambers and Wilson (1968) study, AM performance correlated with retardation and apathy, disorientation, and conceptual disorganization. Empirical evidence indicates, however, that impaired perceptual organization is probably not a trait marker of

292

W. Tschacher et al. / Psychiatry Research 159 (2008) 290–299

schizophrenia spectrum disorder. Results have not been consistent with the assumption of a deficit in all schizophrenia samples. Some studies on perceptual organization have reported that impairments were found only in poor premorbid schizophrenia inpatients, and not in outpatients (Silverstein et al., 1996). Saucer (1958) found no difference between paranoid schizophrenia and normal control subjects with respect to AM perception. Prodromal stages of schizophrenia may even present enhanced perceptual grouping instead of deficiency (Parnas et al., 2001), which resembles phenomenological observations of prodromal gestalt excesses (Conrad, 1958). 1.2. Neurocognitive binding and the dynamical systems framework In contemporary cognitive science, perceptual organization is viewed as a neurocognitive binding process, by which the brain – prior to conscious processing – integrates multiple sensory inputs in order to “bind together” perceptual configurations (Dennett and Kinsbourne, 1992). Synchronous neuronal dynamics presumably underlies such binding processes, whereby selforganized neuronal synchrony is supposed to act as “neuronal glue” by which different brain locations are coordinated (Varela, 1995; Singer and Gray, 1995). Neurocognitive binding has since been embedded in a dynamical systems framework, because gestalt patterns show the characterizing properties of dynamical “attractors”, i.e. asymptotically stable states of a dynamical system (for applications to cognitive systems, see Haken, 1996; Kelso, 1995; Tschacher and Haken, 2007). This can be demonstrated especially in the perception of ambiguous stimuli that gives rise to two or more different attractors. In such cases of multistability, viewers frequently observe transitions between the various attractors; these transitions show the fingerprint of nonlinear phase transitions (mathematical model: Haken et al., 1985; phase transitions in AM: Kruse et al., 1996). The neural basis of transitions is assumed to be neural adaptation (Harris, 1994), a concept consistent with satiation (Koehler, 1944; “depletion of gradients” according to Tschacher and Haken, 2007). Neural adaptation is used to explain the after-effect of motion perception, the effect that viewing of a motion stimulus for a certain period of time leads to an illusory motion in the opposite direction, as soon as the stimulus is discontinued. Based on the dynamical systems model, the stability of attractors is the focus of interest. There are two methods by which this stability can be assessed. First, transition rates can be measured for a fixed period of

time and, second, the response to control parameter variation can be recorded. Stability of perceptual gestalts, therefore, becomes an observable marker of perceptual organization and neurocognitive binding. The present study was conducted to employ these ideas from dynamical systems theory, in an attempt to add to the as yet limited range of neurocognitive measures of schizophrenia (Green et al., 2000). It was specifically intended to clarify the relationship between schizophrenia and perceptual organization. The stabilities of perceptual AM gestalts were defined in three different AM paradigms. Taking research findings on AM, and generally on perceptual organization dysfunction as points of departure, it was expected that AM would be affected in schizophrenia. The primary hypothesis specified that perception of AM depended on the dimensions of psychopathology in a sample of schizophrenia spectrum patients. In this respect it was further expected that the various dimensions would show differential associations with AM perception, thereby explaining the inconsistencies of previous findings. The second, exploratory, hypothesis targeted AM perception as a trait marker of schizophrenia spectrum disorder. Healthy controls were matched to the patient group to investigate general differences between patients, controls and other subgroups, and to study potential specific covariates of AM perception. 2. Methods 2.1. Subjects The study sample of 34 patients comprised 7 (21%) women and 27 (79%) men, with a mean age of 27.9 years (S.D. 7.1). A non-patient control group was matched to the sample with regard to handedness and age (8 women and 26 men; mean age 27.9 years, S.D. 8.0). Patients were recruited from units of the University Hospital of Psychiatry in Bern, Switzerland. A subgroup of n1 = 13 consisted of inpatients admitted to a community-based acute unit (“Soteria Bern”, Ciompi and Hoffmann, 2004). The patients were undergoing psychiatric outpatient treatment in two day hospitals located in Bern. All patients were diagnosed as suffering from schizophrenia spectrum disorder according to the International Classification of Diseases, ICD-10 (F20 schizophrenia, 28; F21 schizotypal disorder, 1; F23 acute psychotic disorder, 2; F25 schizoaffective disorder, 4). The mean dosage of medication in chlorpromazine equivalents prescribed on the day of testing was 267 mg (S.D. 220 mg). Twenty-six patients received atypical neuroleptics, two patients were treated with

W. Tschacher et al. / Psychiatry Research 159 (2008) 290–299

haloperidol decanoate, and six patients were unmedicated. Additional medication with antidepressants (reboxetine or citalopram) or a daily single dose of lorazepame (0.5 mg) was given in three cases. The psychopathological states of 31 patients were determined during standardized clinical interviews (Positive and Negative Syndrome Scale, PANSS; Kay et al., 1987). After inclusion, three patients did not complete the PANSS interviews. Trained staff psychologists, who were not associated with the present project and were naive to its hypotheses, acted as interviewers. The model developed by Lindenmayer et al. (1995) according to which psychopathology as measured by the PANSS is grouped into five factors, i.e. positive, negative, excitement, depression, and cognitive, was used. The positive factor includes symptoms characteristic of florid psychosis such as, for example, delusions, hallucinatory behaviour and unusual thought content. The negative factor includes deficit symptoms of schizophrenia commonly manifested as emotional and social withdrawal. The excitement factor consists of items such as excitement, hostility, tension, and impulsivity. The depression factor consists of the PANSS items depression, anxiety, guilt feelings, somatic concern and preoccupation. And, finally, the cognitive factor incorporates signs of cognitive disorganization, such as conceptual disorganization, disorientation and difficulty in abstract thinking. The mean symptom burden of patients in this study was moderate to low. The Global Assessment of Functioning Scale (GAF, as included in DSM-IV) was implemented to assess the psychological and social levels of functioning in all 34 patients when entering their present course of treatment. The number of previous hospitalizations and together with the patients' ages at the time of their first psychiatric hospitalization were recorded on the basis of case histories. The highest degree of education achieved was operationalised using a 6-point ordinal scale (ranging from 1, high school not completed, to 6, university degree; adapted to the Swiss schooling system). In addition, all participants in the study were required to complete a general reaction time (RT) test. This test consisted of eight runs in which a black square (side length, 6 cm; visual angle, 6.8°) appeared on a screen for a random period of time (mean, 3 s; S.D., 1 s). The patients were instructed to respond to the appearance or disappearance of the square by pressing or releasing the space bar. For study purposes, RT was defined as the response latency averaged across all runs and presses and releases alike. Subjects' characteristics are shown in Table 1. All of the subjects agreed to take part in the present study on the basis of prior informed consent. The study

293

had previously been evaluated and approved by the Canton of Bern Ethics Committee. 2.2. Procedures Three apparent motion (AM) paradigms (see below, a–c) were presented on a 17″ computer screen at a viewing distance of 50 cm. Subjects were instructed to maintain fixation on a cross in the middle of the display. All tests were carried out on working days between 10:00 and 12:00 h. at the testing facility of the hospital's research department. All three paradigms followed the same scheme, whereby one preparatory run (for instructional purposes) preceded the evaluated runs. All data derived from the evaluated runs were pooled per paradigm for ensuing statistical analysis. 2.2.1. Stroboscopic alternative motion (SAM) paradigm The SAM paradigm was introduced by von Schiller (1933) to exemplify AM evoked by stimulus patterns flashed alternately. In addition to the perception of AM – which is reported by all viewers – transitions between different kinds of AM frequently occur (“gestalt flips”). Transitions are perceived spontaneously, i.e. in the absence of environmental or stimulus changes. Thus, the SAM paradigm induces multistability. In the present experiment, the SAM paradigm was realized by showing two stimulus patterns (A and B) alternately and repeatedly without interstimulus intervals (Fig. 1 top). Each SAM stimulus pattern consisted of two black disks of 1 cm in Table 1 Characteristics of control group and patient group Control group Ctrl (n = 34)

Schizophrenia Sz (n = 34)

Variable

M

S.D.

M

S.D.

Age (years) PANSS positive PANSS negative PANSS excitement PANSS depression PANSS cognitive CPE (mg) Education Number of hospitalizations GAF at entry Age at first hospitalization RT (ms) Male/female ratio

27.9

8.0

0 4a

0 2–6b

265 26/8

43

27.9 2.0 1.9 1.5 1.9 1.6 267 3a 3.3 42.6 23.8 324 27/7

7.1 0.9 0.9 0.4 0.6 0.7 220 1–5b 4.5 12.9 5.1 106

Note. PANSS, Positive and Negative Syndrome Scale; CPE, chlorpromazine equivalents; GAF, Global Assessment of Functioning Scale; RT, reaction time; amedian; brange. The following differences are statistically significant (t N 2.0; P b 0.05): Education, Sz b Ctrl (Kruskal-Wallis test). RT, Sz N Ctrl.

294

W. Tschacher et al. / Psychiatry Research 159 (2008) 290–299

Fig. 1. Top panel, the stimulus patterns (A, B) presented alternately in the SAM paradigm. Bottom panel, the stimulus patterns (A, B) presented alternately in the CAM paradigm.

diameter (visual angle, 1.1°) shown against a white background. The disks were placed at the corners of an imaginary rectangle with horizontal sides of 7.7 cm (visual angle, 8.8°) and vertical sides of 8.7 cm (visual angle, 9.9°). The aspect ratio was, therefore, 7.7:8.7 ≈ 0.89. Each presentation of the stimulus pattern A or B lasted 500 ms (flashing frequency, 2 Hz). One run of successive presentations A–B–A–B–A–… lasted 60 s. Three such runs were presented to each subject. Prior to the first run, all subjects (who were collectively naive to AM phenomena) were shown the experiment and informed that three different perceptions were possible, i.e. vertical AP, horizontal AP, and the rarer circular AP (either clockwise or counterclockwise). Subjects were then instructed to press different keys depending on the AM gestalt they actually perceived, and to keep the key pressed for as long as they perceived that specific gestalt. In this way, the various perceptions and durations of AM gestalts were recorded throughout the experiment. For each subject, the number of SAM transitions was pooled across all three runs to assess the global stability of spontaneously arising SAM gestalt perceptions. This variable was labelled “SAM transition rate”. 2.2.2. SAM paradigm with aspect ratio variation This paradigm was developed on the basis of the SAM paradigm described above. It has been established that the probability of perceiving either vertical or horizontal AM closely depends on the aspect ratio of the display (Kruse et al., 1996). Hence, the aspect ratio constitutes a control parameter of SAM perception.

Large aspect ratios, i.e. the horizontal sides of the imaginary rectangle in Fig. 1 (top panel) exceed its vertical sides, are associated with a higher probability of perceiving vertical AM. When the aspect ratio is reduced to zero, a spontaneous gestalt flip from vertical AM to horizontal AM occurs. In some cases, the gestalt flip occurs from vertical AM to circular AM. Each run of this paradigm started with an aspect ratio of 2.1 (horizontal extension, 18.6 cm; vertical extension, 9 cm; visual angles 20.4° and 10.2°, respectively). The aspect ratio was then gradually reduced to zero by keeping the vertical extension constant while reducing the horizontal extension step by step with each successive frame. Each run lasted 30 s (i.e. 60 single frames were presented with a flashing frequency of 2 Hz as in the SAM paradigm). A total of six runs were presented. Subjects recorded the times at which gestalt flips occurred by key presses. This paradigm was represented by the time (ms) that elapsed until vertical AM switched (with decreasing aspect ratio) to any other perceived AM gestalt. The variable “time to SAM transition” was defined as the average of all six runs per participant. 2.2.3. Circular apparent motion (CAM) paradigm A circular configuration of six disks was presented as in Fig. 1 (bottom). The disks were 1 cm in diameter (visual angle, 1.1°). The diameter of the configuration was 8.8 cm (visual angle, 10°), and the distance between two juxtaposed disks was 4.4 cm (visual angle, 5°). Frame B is identical to A after rotation by an angle of 30°. Frames A and B in Fig. 1 (bottom panel) were flashed alternately with a frequency of 2 Hz. Each trial lasted 60 s. The CAM paradigm provokes spontaneous gestalt transitions in the same way as the SAM paradigm. Subjects were instructed to press three different keys for three different AM gestalts, two different keys for clockwise and counterclockwise circular AM, and one key when perceiving disks moving back and forth (oscillatory AM). The variable “CAM transition rate” was defined as the number of gestalt transitions pooled across the three runs completed by each participant. 2.3. Statistical analyses Multivariate analysis of variance (MANOVA) and multiple regression analyses were performed to estimate if, and in what way, AM perception was linked to dimensions of psychopathology (primary hypothesis). In the regression platform, each of the three AM variables was treated as a dependent measure and the five PANSS psychopathology factors as predictors. In addition to whole model regression analysis, backward

W. Tschacher et al. / Psychiatry Research 159 (2008) 290–299

stepwise regression was applied to determine the most parsimonious model by which AM could be predicted. AM perception in the patient group and the control group was compared by logistic regression analysis (secondary hypothesis). To identify subgroups in the overall sample that were homogeneous as to clinical and demographic variables, latent class analysis (LCA) was used. LCA is a model-based clustering technique that provides information on the selection of an appropriate number of clusters (Magidson and Vermunt, 2001) within a sample. LCA assumes that hidden structures (e.g. sample subgroups) in a data set can be described by means of one or more unchanging discrete latent (unobserved) variables that based on observed, manifest variables divide the sample into mutually exclusive and exhaustive latent classes. Manifest variables (such as RT in a cognitive test) serve as indicators of the latent variables. A standard exploratory procedure for fitting LCA models is to start with a one-class (independence) model, then increase the number of latent classes until the solution is found that best fits the data. The SchwarzBayesian information criterion (BIC) and likelihoodratio test statistics were applied to identify the best model (Sakamoto et al., 1986). The Latent Gold software package was used for these analyses (Vermunt and Magidson, 2000). Finally, a path model was computed to summarize the relationships between the latent classes of the sample and their performance in the AM paradigms using the AMOS software package (Arbuckle, 2003).

295

Table 2 Summary of the whole model ANOVA tests and stepwise regression analyses for psychopathology factors predicting apparent motion Multiple Stepwise regression regression (backward) whole model AM variable

R2

F

R2

PANSS predictor (beta weight) 0.29 2.08 0.22 4.01 ⁎ positive (0.40) depression (− 0.33) 0.48 4.61 ⁎⁎ 0.37 17.2 ⁎⁎ negative (0.61)

SAM transition rate Time to SAM transition CAM 0.37 2.97 ⁎ transition rate

F

0.36 5.05 ⁎⁎ cognitive (0.45) excitement (−0.43) negative (− 0.38)

Note. AM, apparent motion; SAM, stroboscopic apparent motion; CAM, circular apparent motion. n = 31. ⁎ P b 0.05. ⁎⁎ P b 0.01.

associated with psychopathology. The association between SAM transition rates and psychopathology was weaker, however, as only the stepwise regression model was statistically significant. As expected, the signs of predictors (beta weights) showed that psychopathology dimensions contributed differentially to gestalt perception. The positive factor and the cognitive factor were both linked to reduced gestalt stability (i.e. increased transition rates), whereas the negative, excitement and depression factors were all related to the stability of gestalt perception (i.e. decreased transition rates and delayed time to SAM transition).

3. Results 3.2. Differences between subgroups 3.1. Association of apparent motion perception with psychopathology In an encompassing MANOVA test, the hypothesis of a general relationship between all AM variables (as dependent responses) and all psychopathology variables (the independent variables) was assessed. The overall test was significant (F(15, 63.9) = 2.46, P b 0.01; Wilk's lambda), indicating a clear association between AM and psychopathology. The results of multiple regression analyses are provided in Table 2. It was found that significant regression models were derived for all paradigms. The highest proportion of explained variance existed in the CAM task and the varied SAM task, in which up to 48% of the total variance was ascribable to psychopathology predictors. Hence, the performance in CAM and the timing of gestalt transitions in the SAM paradigm with aspect ratio variation were found to be strongly

No significant differences were found between participants with schizophrenia spectrum disorder and the control group. A whole model test (logistic regression analysis of AM variables with ‘group’ as the dependent variable) yielded an insignificant result (N = 68; df = 3; χ2 = 1.6; P = 0.65). It was found that the patient group was heterogeneous with regard to AM perception. The performance of the day hospital patients (outpatients, n2 = 21) in this sample was different from that of control subjects and inpatients alike (n1 = 13). Logistic regression analysis of AM variables with ‘subgroup’ (i.e. outpatients vs. inpatients vs. control subjects) as a dependent variable was significant (N = 68; df = 6; χ2 = 25.3; P b 0.001), explaining 18% of variance. In all three AM paradigms, outpatients consistently displayed higher levels of gestalt stability than the control group, and inpatients showed lower stability than the control group. AM

296

W. Tschacher et al. / Psychiatry Research 159 (2008) 290–299

perception was not significantly related to chlorpromazine equivalents in the patient group (Pearson correlations, − 0.17 with “SAM transition rate”; − 0.06 with “time to SAM transition”; − 0.03 with “CAM transition rate”). The latent class analysis (LCA) based on participants' manifest characteristics (variables age, sex, educational level, mean RT) revealed the existence of two subgroups within the overall sample (Fig. 2). This two-class solution yielded the best fit to the sample data both according to the information criterion BIC and the likelihood ratio test. Class A of the two-class solution consisted of 53% (n = 36) of the sample. Individuals in this latent class were more likely to be male, were younger and showed shorter RT than those in latent class B. Class B subjects (n = 32) had higher mean age and longer RT. Classes were not congruent with the distinction of patient versus control group; both classes contained members of the healthy control group. Patients with differing treatment status were distributed unevenly between latent classes: class A included 11 (85%) of 13 inpatients, whereas class B included 17 (81%) of 21 day hospital outpatients. The path model depicted in Fig. 2 showed a significant relationship between class membership and gestalt perception — class A subjects were characterized by low levels of gestalt stability (i.e. high frequencies of AM transitions and short time to SAM transition); class B subjects showed a contrasting pattern of performance in the three AM paradigms, namely low numbers of

transitions and delayed SAM gestalt flips (logistic regression analysis of AM variables with ‘class’ as dependent variable: N = 68; df = 3; χ2 = 12.2; P b 0.01). 4. Discussion During early stages of cognition, mechanisms of neurocognitive binding organize the informational input prior to conscious processing. This ‘gestalt perception’ is of considerable importance to cognition, in that it sets the stage for subsequent steps, such as the allocation of attention, memory processes, and executive functioning. From a first-person perspective, gestalt formation organizes the experienced world, rendering it reliable and familiar, while categorizing information in a way that supports recognition and action. According to our general hypothesis, this process is affected in schizophrenia spectrum disorders. In the same vein, Parnas et al. (2001) recognized in schizophrenia “a deficiency in the automatic pre-reflective intentionality” (p. 172). Three different paradigms were implemented to test apparent motion (AM) gestalts in patients with schizophrenia spectrum disorder. It was found that AM perception was strongly associated with psychopathology factors derived from PANSS interviews. The symptoms had differential links with perception; both the positive syndrome factor and the cognitive (disorganization) factor were related to increased switching between gestalt perceptions. Patients exhibiting positive psychosis and cognitive disorganization thus experienced reduced

Fig. 2. Schema of latent class analysis (left) and path model (right, with arrows and standardized path coefficients).

W. Tschacher et al. / Psychiatry Research 159 (2008) 290–299

gestalt stabilities. The other dimensions of schizophrenia psychopathology (the negative syndrome factor, the excitement factor, the depression factor) had opposing regression weights, connected to lowered gestalt transition rates. In summary, these findings contribute to the body of literature on perceptual organization in schizophrenia and, after a lag of several decades, serve to once again extend schizophrenia research to the field of AM phenomena. Saucer concluded in 1958 that AM measures “may be a useful diagnostic tool but may be better suited to research”, due to the complex electronic apparatus needed in those years. Computer-based AM tools are easily available nowadays; thus, it appears promising to pursue Saucer's thread. Secondly, by including a matched control group, the hypothesis whereby deviation in AM gestalt processing constitutes a trait marker of schizophrenia spectrum disorders, was tested. Not unexpectedly, such specificity assumptions were not corroborated by the data, in that no significant differences were evident when comparing patients and control participants. At the same time, two latent classes were found in the overall sample on the basis of demographic and reaction time measures. These two classes, however, were significantly discriminative with respect to gestalt performance; latent class analysis thus sorted the sample in a meaningful way. Class A members who had low levels of AM stability comprised the majority of inpatients and control subjects; such participants were generally young, male, with short RT. Class B typically contained outpatients and some control subjects; participants in class B were older and showed longer RT. Thus, AM gestalt alterations were not found to be specific to schizophrenia spectrum disorder per se. Results may be discussed using the concept of state-dependence, which deals with the process character of schizophrenia (Exner et al., 2006; Tschacher, 2004). Altered gestalt perception is specific in the sense that the cognitive function of pre-attentional neurocognitive binding is focally affected by the disorder; it is stage-specific, however, because both the up- and down-regulation of neurocognitive binding is found depending on the stage of the disorder. Enhancement of gestalt perception is found in prodromal (Parnas et al., 2001) and incipient schizophrenia, which was characterized by Conrad (1958) as the trema stage. Behavior of class A patients is consistent with reports of preserved gestalt functioning in schizophrenia (Chey and Holzman, 1997; Herzog et al., 2004). Reduction of gestalt perception is reliably found in chronic schizophrenia, especially with poor premorbid adaptation (Silverstein and Uhlhaas, 2004). We therefore posit AM gestalt processing as a core “stage

297

marker” in schizophrenia spectrum disorders with potential usefulness as a diagnostic and predictive tool. A limiting aspect of the present study may lie with the recruitment of young patients characterized by relatively low symptom burdens, a particular which may have reduced the testing power owing to floor effects. This sample showed relatively good premorbid adaptation. It is therefore possible that the contribution of the cognitive factor in the present sample was underestimated. A further caveat concerns the relationship between psychopathology and AM perception. Which of the two is the causal foundation to the other? This important issue could not be explored in a crosssectional design. One would be inclined to assume that neurocognitive binding, assessed via AM perception, is responsible for ensuing psychotic and general symptoms. Symptoms would thus be understood as expressions of neurocognitive dysfunction. Alternatively, both symptoms and AM perception may reflect a disturbance of a core algorithm of cognitive coordination (Phillips and Silverstein, 2003). It was beyond the scope of the present study to pursue this link and the possible causal feedbacks between psychopathology and neurocognitive binding. The stage specificity postulation underscores the general notion that schizophrenia spectrum disorders are processes. Rather than searching for the trait marker of schizophrenia, symptoms and underlying neurocognition alike should be viewed as nonlinear functions of time in a dynamical disease framework (Pezard et al., 1996; Tschacher et al., 1997). This dynamical framework, paired with the exploitation of longitudinal data and time series analyses (Tschacher and Kupper, 2002), afford avenues for future research. A consideration of the temporal dimension would most especially help in the elucidation of the causal relationships between neurocognitive binding and symptoms. Acknowledgments This study was in part supported by the Swiss National Foundation grant 32-55954. The authors thank the staff and directors of Soteria Bern and two day hospitals. We are grateful for the help provided by Daniela Schuler and Simon Grossmann. References Arbuckle, J.L., 2003. Amos 5.0 Update to the Amos User's Guide. Smallwaters Corporation, Chicago, IL. Brand, A., Kopmann, S., Marbach, S., Heinze, M., Herzog, M.H., 2005. Intact and deficient feature fusion in schizophrenia. European Archives of Psychiatry and Clinical Neuroscience 255, 413–418.

298

W. Tschacher et al. / Psychiatry Research 159 (2008) 290–299

Chambers, J.L., Wilson, W.T., 1968. Perception of apparent motion and degree of mental pathology. Perceptual and Motor Skills 26, 855–861. Chen, Y., Nakayama, K., Levy, D.L., Matthysse, S., Holzman, P.S., 1999. Psychophysical isolation of a motion-processing deficit in schizophrenics and their relatives and its association with impaired smooth pursuit. Proceedings of the National Academy of Sciences of the United States of America 96, 4724–4729. Chey, J., Holzman, P.S., 1997. Perceptual organization in schizophrenia: utilization of the gestalt principles. Journal of Abnormal Psychology 106, 530–538. Ciompi, L., Hoffmann, H., 2004. Soteria Berne — an innovative milieu therapeutic approach to acute schizophrenia based on the concept of affect-logic. World Psychiatry 3, 140–146. Conrad, K., 1958. Die beginnende Schizophrenie [Early Schizophrenia]. Thieme, Stuttgart. Dennett, D.C., Kinsbourne, M., 1992. Time and the observer: the where and when of consciousness in the brain. Behavioral and Brain Sciences 15, 183–247. Doniger, G.M., Silipo, G., Rabinowicz, E.F., Snodgrass, J.G., Javitt, D.C., 2001. Impaired sensory processing as a basis for object-recognition deficits in schizophrenia. American Journal of Psychiatry 158, 1818–1826. Exner, C., Boucsein, K., Degner, D., Irle, E., 2006. State-dependent implicit learning deficit in schizophrenia: evidence from 20-month follow-up. Psychiatry Research 142, 39–52. Goebel, R., Khorram-Sefat, D., Muckli, L., Hacker, H., Singer, W., 1998. The constructive nature of vision: direct evidence from functional magnetic resonance imaging studies of apparent motion and motion imagery. European Journal of Neuroscience 10, 1563–1573. Goodarzi, M.A., Wykes, T., Hemsley, D.R., 2000. Cerebral lateralization of global–local processing in people with schizotypy. Schizophrenia Research 45, 115–121. Green, M.F., Kern, R.S., Braff, D.L., Mintz, J., 2000. Neurocognitive deficits and functional outcome in schizophrenia: are we measuring the “right stuff”? Schizophrenia Bulletin 26, 119–136. Haken, H., 1996. Principles of Brain Functioning. Springer, Berlin. Haken, H., Kelso, J.S., Bunz, H., 1985. A theoretical model of phase transitions in human hand movements. Biological Cybernetics 51, 347–356. Harris, J., 1994. The duration of the movement aftereffect as an index of psychiatric illness. Perception 23, 1145–1153. Herzog, M.H., Kopmann, S., Brand, A., 2004. Intact figure-ground segmentation in schizophrenia. Psychiatry Research 129, 55–63. Kay, S.R., Fiszbein, A., Opler, L.A., 1987. The positive and negative syndrome scale (PANSS) for schizophrenia. Schizophrenia Bulletin 13, 261–276. Kelso, J.S., 1995. Dynamic Patterns: The Self-Organization of Brain and Behavior. MIT Press, Cambridge. Kéri, S., Kelemen, O., Benedek, G., Janka, Z., 2005. Lateral interactions in the visual cortex of patients with schizophrenia and bipolar disorder. Psychological Medicine 35, 1043–1051. Koehler, W., 1944. New facts in visual perception. Transactions of the New York Academy of Sciences 7, 39–42. Kruse, P., Carmesin, H.-O., Pahlke, L., Strüber, D., Stadler, M., 1996. Continuous phase transitions in the perception of multistable visual patterns. Biological Cybernetics 75, 321–330. Lindenmayer, J.-P., Grochowski, S., Hyman, R.B., 1995. Five factor model of schizophrenia: replication across samples. Schizophrenia Research 14, 229–234. Magidson, J., Vermunt, J.K., 2001. Latent class factor and cluster models, bi-plots and related graphical displays. Sociological Methodology 31, 223–264.

Matussek, P., 1952. Untersuchungen über die Wahrnehmung. 1. Mitteilung. Archive für Psychiatrie und Zeitschrift für Neurologie 180, 279–319. [English version in: Cutting, J., Sheppard, M. (Eds.), 1987. Clinical Roots of the Schizophrenia Concept. Translations of Seminal European Contributions on Schizophrenia. Cambridge University Press, Cambridge, pp. 87–103.] Mitroff, S.R., Scholl, B.J., 2005. Forming and updating object representations without awareness: evidence from motion-induced blindness. Vision Research 45, 961–967. Parnas, J., Vianin, P., Saebye, D., Jansson, L., Volmer-Larsen, A., Bovet, P., 2001. Visual binding abilities in the initial and advanced stages of schizophrenia. Acta Psychiatrica Scandinavica 103, 171–180. Pezard, L., Nandrino, J.-L., Renault, B., Massioui, F.E., Allilaire, J.-F., Varela, F.J., 1996. Depression as a dynamical disease. Biological Psychiatry 39, 991–999. Phillips, W.A., Silverstein, S.M., 2003. Convergence of biological and psychological perspectives on cognitive coordination in schizophrenia. Behavioral and Brain Sciences 26, 65–138. Place, E.J., Gilmore, G.C., 1980. Perceptual organization in schizophrenia. Journal of Abnormal Psychology 89, 409–418. Sakamoto, Y., Ishiguro, M., Kitagawa, G., Akaike, H., 1986. Information Criterion Statistics. Series: Mathematics and its Applications, vol. 1. Springer, New York. Saucer, R.T., 1958. A further study of the perception of apparent motion by schizophrenics. Journal of Consulting Psychology 22, 256–258. Saucer, R.T., Deabler, H.L., 1956. Perception of apparent motion in organics and schizophrenics. Journal of Consulting Psychology 20, 385–389. Silverstein, S.M., Uhlhaas, P.J., 2004. Gestalt psychology: the forgotten paradigm in abnormal psychology. American Journal of Psychology 117, 259–277. Silverstein, S.M., Knight, R.A., Schwarzkopf, S.B., West, L.L., Osborn, L.M., Kamin, D., 1996. Stimulus configuration and context effects in perceptual organization in schizophrenia. Journal of Abnormal Psychology 105, 410–420. Silverstein, S.M., Kovacs, I., Corry, R., Valone, C., 2000. Perceptual organization, the disorganization syndrome, and context processing in chronic schizophrenia. Schizophrenia Research 43, 11–20. Singer, W., Gray, C.M., 1995. Visual feature integration and the temporal correlation hypothesis. Annual Review of Neuroscience 18, 555–586. Tschacher, W., 2004. How specific is the gestalt-informed approach to schizophrenia? Gestalt Theory 26, 335–344. Tschacher, W., Haken, H., 2007. Intentionality in Non-Equilibrium Systems? The Functional Aspects of Self-Organized Pattern Formation. New Ideas in Psychology 25, 1–15. Tschacher, W., Kupper, Z., 2002. Time series models of symptoms in schizophrenia. Psychiatry Research 113, 127–137. Tschacher, W., Scheier, C., Hashimoto, Y., 1997. Dynamical analysis of schizophrenia courses. Biological Psychiatry 41, 428–437. Tschacher, W., Schuler, D., Junghan, U., 2006. Reduced perception of the motion-induced blindness illusion in schizophrenia. Schizophrenia Research 81, 261–267. Uhlhaas, P.J., Silverstein, S.M., 2003. The continuing relevance of Gestalt psychology for an understanding of schizophrenia. Gestalt Theory 25, 215–240. Uhlhaas, P.J., Silverstein, S.M., 2005. Perceptual organization in schizophrenia spectrum disorders: empirical research and theoretical implications. Psychological Bulletin 131, 618–632. Varela, F.J., 1995. Resonant cell assemblies: a new approach to cognitive function and neuronal synchrony. Biological Research 28, 81–95.

W. Tschacher et al. / Psychiatry Research 159 (2008) 290–299 Vermunt, J.K., Magidson, J., 2000. Latent GOLD's User's Guide. Statistical Innovations Inc., Boston. von Holst, E., Mittelstaedt, H., 1950. Das Reafferenzprinzip. Wechselwirkung zwischen Zentralnervensystem und Peripherie [The reafference principle: mutual relationship between central nervous system and periphery]. Naturwissenschaften 37, 464–476.

299

von Schiller, P., 1933. Stroboskopische Alternativversuche [Stroboscopic alternation trials]. Psychologische Forschung 17, 179–214. Wertheimer, M., 1912. Experimentelle Studien über das Sehen von Bewegungen [Experimental studies on visual perception of motion]. Zeitschrift für Psychologie 61, 161–265.