go to content


Vadim S. Rotenberg   vadir@post.tau.ac.il, J. Hadjez, P. Indursky, T. Martin, Y. Barak, Y.Gutman Abarbanel Mental Health Center, Sackler School of Medicine, Tel Aviv University, Israel

Homeostasis 1997, 38: 97-102.

Polysomnography of 3 consecutive nights was performed in 20 drug treated chronic schizophrenic patients in parallel with scoring of positive and negative symptoms (PANSS). Eye movement density - estimated separately for each cycle - was lower in patients with a relative prevalence of positive symptoms in comparison with patients with predominantly negative symptoms. Eye movement density correlated negatively with the score of positive symptoms. The results are discussed in the context of the Search Activity concept, by taking into consideration the regulation of the saccadic eye movements in sleep and wakefulness by the right hemisphere, which is sensitive to neuroleptic treatment.

Key words: Schizophrenia positive vs. negative; eye movements; REM sleep.

Eye movement density is one of the main phasic REM sleep variables. Previous investigations have shown that eye movement (EM) density correlates with dream content in healthy subjects and in mentally ill patients (Herman, et al., 1984; Rotenberg, 1988). Depression is characterized by increased EM density in the first cycles (Foster, et al., 1976; Reynolds III, et al., 1988), while in healthy subjects EM density increases from the first to the subsequent cycles.

Data regarding EM density and EM distribution between cycles in schizophrenia is less definite. Most studies of REM sleep in schizophrenia have been performed on drug naive patients or soon after withdrawal of neuroleptic drugs. The results of these studies may be summarized as follows:

1. EM density was increased during neuroleptic treatment in comparison with neuroleptic withdrawal (Neylan et al., 1992).

2. Immediately after drug withdrawal, EM density was higher in comparison to drug naive patients and patients with a long-lasting drug discontinuation (Tandon, et al., 1992).

3. The increased severity of psychotic symptoms after drug withdrawal correlated with the decrease of REM activity and EM density (Neylan, et al., 1992).

4. EM density correlated positively with the hallucinatory behavior (Feinberg, 1971; Benson and Zarcone, 1993), a finding which seems to be in contradiction with the above mentioned results of Neylan, et al. (1992).

We have not found in the literature a systematic study of the relationships between EM density and positive and negative symptoms in schizophrenic patients maintained on stable neuroleptic treatment.


The sample consisted of 20 patients with chronic schizophrenia, 13 men and 7 women, at a mean age of 45.8 ( s.d. standard deviation- 6.9) years. Duration of illness ranged from 11 to 39 years (mean duration 21.6, s.d. 4.7 years). Clinical evaluation of the disease was performed using the structured clinical interview and the diagnosis was established according to DSM-IV criteria: 11 patients met the criteria for Paranoid Schizophrenia; 3 for Disorganized Schizophrenia; 5 for Residual Schizophrenia; 1 for Schizoaffective Disorder. PANSS ratings (Kay, et al., 1987) were performed in order to evaluate the severity of positive and negative signs. Patients were free from somatic medical disorders.

All patients were on neuroleptic medication during the study, and their treatment had not changed for 60 days prior to polysomnography. Hypnotic medication was discontinued at least two weeks prior to polysomnography.

Sleep data was collected from all patients on 3 consecutive nights. Lights were usually turned off between 10 and 10:30 PM, and turned on between 6 and 6:30 A.M., according to the usual ward schedule. Electroencephalogram, submental electromyogram, and electroocculograms were recorded at a low filter setting of 0.3 Hz.

Polysomnograms of 60 nights were analyzed according to international criteria (Rechtschaffen and Kales, 1968). Only EM density data are reported in this paper. Eye movements in REM sleep were counted visually. The criterion for rapid EM detection was a minimum of 25microV excursion (Benson and Zarcone, 1993). The EM density was estimated as an averaged EM frequency during 1 minute of stage REM sleep in each cycle. Mean EM density means were calculated for nights, subjects, and for groups of patients. In addition, mean EM density was calculated in patients with the relatively high level of positive/negative symptoms (higher than mean data for the whole group plus standard deviation), and in patients with relatively low level of positive/negative symptoms (less than the mean data for the whole group minus standard deviation).

Student's t-test was used for the comparison of positive and negative symptoms scores or EM density values in subgroups of patients. Product moment correlation was used for the estimation of relationships between positive vs. negative symptoms and EM density.


In patients with a low EM density, the mean score of positive symptoms was significantly higher than in patients with a high EM density (p <0.001); the scores of negative symptoms were similar in both subgroups. In patients with a low score of the positive symptoms, mean EM density was significantly higher than in patients with a high score (p <0.05); EM density was similar in patients with high and low scores of the negative symptoms. (Tab. 1)

Table 1
  EM density POSITIVE symptoms NEGATIVE symptoms
All patients 5.3 (s.d.2.8) 17.1 (s.d. 7.5) 18.7(s.d. 5.2)


25.2 (s.d. 3.5) 33.2 (s.d. 4.7)


12.6(s.d. 1.8)

31.0(s.d. 3.0)


5.25 (s.d. 0.72)



4.0(s.d. 1.1)



4.9(s.d. 1.0)




4.3 (s.d. 3.8)



In patients with the ratio of positive to negative symptoms higher than 0.6, the EM density was lower than in patients with this ratio lower than 0.4, for all sleep cycles. In the first and the third cycle this difference was significant: mean density values were 2.6 (sd 3.0) vs. 8.1 (sd 5.1) for the first cycle, and 3.6 (sd 2.3) vs. 6.6 (sd 2.5) for the third one. In subjects with the positive/negative ratio higher that 0.6, EM density in the first cycle was lower than in other cycles. In subjects with the ratio lower than 0.4, EM density achieved maximum values during the first cycle, in other cycles it was lower and almost equal, thus resembling EM distribution in depressed patients (Foster, et al, 1976). A significant (p <0.05) negative correlation was found between the score of the positive symptoms and EM density in the first and third cycles (-.45; -.50).


To summarize the main results of the present study: 1. Positive symptoms were higher in schizophrenic patients with a lower EM density. 2. Subjects with the ratio between positive and negative symptoms higher than 0.6, had EM density in the 1st and 3rd cycles significantly lower than patients with this ratio lower than 0.4.

These findings are in agreement with Neylan, et al., (1992), reporting an increased severity of psychosis after neuroleptic withdrawal simultaneously with a decrease in EM density. Our results correspond also with Tandon, et al., (1992), reporting that schizophrenic patients who are drug free only for 2 to 4 weeks (when the effect of neuroleptics is still present) display higher EM density in comparison with schizophrenic patients who are drug free for a longer time or drug naive. Tandon, et al., have not presented data about positive and negative symptoms after drug withdrawal. It is, however, not rare that the withdrawal of neuroleptics causes the exacerbation of psychotic (positive) symptoms. A comparison of our data with the results of both these studies suggests that positive symptoms correlate negatively with EM density during neuroleptic treatment as well as after treatment discontinuation. These results were predicted theoretically (Rotenberg, 1994), on the basis of the Search Activity concept (Rotenberg, 1984; 1993; 1994).

By Search Activity (SA), we understand activity oriented to change the situation (or subject's attitudes to it) in the absence of the precise prediction of the outcome of such activity, but considering the accomplished outcomes at all stages of activity. This describes a type of feedback cycle related to problem solving. SA is a component of self-stimulation in animals and creative behavior in humans, as well as exploratory and active defense behavior in both species. SA is absent in free-floating neurotic anxiety and depression in humans, in freezing in animals (renunciation of search), as well as in panicky behavior, learned helplessness, and stereotyped behavior in both species (Rotenberg, 1984). Any form of behavior that contains SA increases body resistance to various forms of artificial and natural pathology while renunciation of search decreases body resistance and accelerates the development of pathology (Rotenberg and Arshavsky, 1979; Rotenberg, 1996).

It was shown that SA decreases and renunciation of search increases REM sleep requirement because the function of REM sleep is to compensate in dreams for the deficit of SA in the previous state of wakefulness, and to restore SA in the subsequent wakefulness (for details, see Rotenberg, 1984, 1993). Thus, if SA is high in wakefulness, the requirement for REM sleep and its activity is low. EM density in healthy subjects correlates with the active participation of the dreamer in dream events (Rotenberg, 1993). If SA during wakefulness is already high, there is no need for high EM density in REM sleep.

Positive symptoms, e.g., paranoid ideation and hallucinations, correspond to the definition of search activity: they display active behavior without definite probability forecast. A recent study (McGuire, et al., 1993; 1996) showed that auditory hallucinations are associated with increased metabolic activity of brain centers for inner speech, i.e., that represent active verbalization. In the state of positive schizophrenia, a subject never can be sure about future events" in the artificial world he is building for himself, or about outcomes of his interactions with this world. At the same time, the subject remains highly sensitive to all events and outcomes that are related to his positive symptoms. Certainly, this form of search activity" is misdirected and maladaptive from the social point of view, but misdirection does not contradict the definition of search activity. Thus, the high level of pathological SA determines negative correlation between positive symptoms and EM density.

Our results seem to be in direct contradiction with the results of some other investigators. Feinberg (1971), and Beanson and Zarcone (1993), found that EM density correlated positively with hallucinatory behavior. How can this contradiction be explained? One difference between our investigations and the above mentioned, is that our patients have been on continuous neuroleptic treatment, while theirs have been drug free during sleep investigation. We suggest that this can be an important factor, and to facilitate the discussion we need to present a short review of the interrelationships of brain hemisphere asymmetry, eye movement activity in wakefulness and dream states, neuroleptic treatment, and positive vs. negative symptoms.

Eye movements in REM sleep are saccadic scans of targets in the dream scene (Hong, et al., 1955). The right hemisphere is specialized for saccadic eye movements in both REM sleep and wakefulness. The superiority of the right hemisphere in oculomotor control was shown by Bracewel, et al., (1990). Patients with right hemisphere damage and left visual hemiattention showed a deficit in leftward-directed eye movements in REM sleep (Doricchi, et al., 1991). Neuroleptic-naive schizophrenic patients display left hemisphere dysfunction manifested in an impaired orienting to visual targets in the right visual field (Posner, et al., 1988). We suggest that such left-hemisphere dysfunction corresponds to the predomination of positive symptoms, which are usually based on left hemisphere activity (Gur, et al., 1978; Flor-Henry, 1983) and grasps subjects attention. This assumption in confirmed by data that more symptomatic schizophrenic patients show a tactile-kinesthetic hemineglect of the right side of space, while the less symptomatic, of the left side (Harvey, 1993). In acutely psychotic patients who were never medicated (and as a result, presented positive symptoms), there was a significant right-sided hemineglect which then shifted to the left with medication (Tower and Flor-Henry, 1989).

The preceding findings lead to the following conclusions:

1. Positive Symptoms are related predominantly to the dysfunction (hyperactivity and occupation with hallucinations) of the left hemisphere, which is not involved in the regulation of oculomotor activity.

2. Negative Symptoms are related predominantly to the dysfunction (hypofunction) of the right hemisphere (Cutting, 1990; Rotenberg, 1994), which is responsible for oculomotor activity in wakefulness and dreams.

3. Neuroleptic Treatment, which improves the activity of the left hemisphere and abolishes the positive symptoms, does not treat the primary negative symptoms of the deficit syndrome (Kirkpatrick and Carpenter, 1995) and does not restore the functional activity of the right hemisphere. Moreover, with traditional neuroleptics, treatment may suppress the right hemisphere activity (Tower, Flor-Henry, 1989). As a result, under neuroleptic treatment, the normal regulation of EM activity might be disturbed.

This may be the cause of the positive correlation between EM density and hallucinations in non-medicated patients. Hallucinations may display an indirect sign of the relatively less disturbed right hemisphere functions, compared to schizophrenia with the predomination of negative symptoms without hallucinations. Under traditional neuroleptic treatment, the right hemisphere becomes additionally functionally suppressed, and prevents the correlation between rapid eye movements and positive symptoms, even if they are still present. In non-treated patients, the more predominant are positive symptoms vs. negative, the more involved in the process is the left hemisphere, and the more functionally sufficient is the right hemisphere. This can explain the correlation between positive symptoms and EM density.

To sum up the discussion, two factors may interfere in the process of EM regulation in schizophrenia. The first is the presence or absence of the positive symptoms, which in general decreases the requirement for REM sleep activity. The second factor is the functional sufficiency of the right hemisphere. In patients with positive symptoms and without neuroleptic treatment, the right hemisphere is relatively more sufficient than in patients with predominantly negative symptoms. As a result, the second factor can partly neutralize the first: increased Search Activity decreases the requirement for REM sleep activity, but at the same time right hemisphere preserves the regulation of EM activity. In patients with predominantly negative symptoms, the requirement for REM activity is high, but the functional insufficiency of the right hemisphere disturbs the regulation of EM activity in target-oriented attention and makes the satisfaction of this requirement impossible. This can explain the absence of positive correlation
between EM density and negative symptoms. And finally, in treated schizophrenic patients, positive symptoms(misdirected search activity) may correlate negatively with EM activity while hallucinations do not correlate positively with EM because neuroleptic treatment prevents regulation and activation of eye movements.


Benson E. and Zarcone V. P. (1993): Rapid eye movement sleep eye movements in schizophrenia and depression. Arch. Gener. Psychiat. 50: 474-482.

Brace well R. M., Husain M. and Stein J. F. (1990): Specialization of the right hemisphere for visuomotor control. Neuropsychologia, 28: 763-775.

Cutting J. (1990): The right cerebral hemisphere and psychiatric disorders. Oxford University Press.

Doricchi F., Guariglia C., Paolucci S., and Pizzamiglio L. (1991): Disappearance of leftward eye movements during sleep in left visual hemiattention. Neuroreport, 2: 285-288.

Feinberg I., Koresko R. L. and Gottlieh F. (1965): Further observations on electrophysiological sleep patterns in schizophrenia. Comprehensive Psychiatry, 6: 21-24.

Flor-Henry P. (1983): Cerebral basis of psychopathology. Boston, Bristol, London.

Foster F. G., Kupfer D. Y., Coble P. and McParland R. J. (1976): Rapid eye movement sleep density. Arch. Gener. Psychiat. 33: 119-123.

Gur R. E. (1978): Left hemisphere dysfunction and left hemisphere overactivation in schizophrenia. J. Abnorm. Psychol. 87: 226-238.

Harvey S. A., Nelson E., Haller J. W. and Early T. S. (1993): Lateralized attentional abnormality in schizophrenia is correlated with severity of symptoms. Biolog. Psychiatry, 33: 93-99.
Herman J. H., Erman M., Boys R., Peiser L., Taylor M. E., Roffwarg H. (1984): Evidence for a directional correspondence between eye movements and dream imagery in REM sleep. Sleep, 7: 52-63.
Hong Ch. Ch.-H., Gillin J. Ch., Dow B. M., Wu J. and Buchsbaum M. S. (1995): Localized and lateralized cerebral glucose metabolism associated with eye movements during REM sleep and wakefulness: A positron emission tomography (PET) study. Sleep, 18: 579-580.

Kay S. R., Fiszbein A. and Opler L. A. (1987): The positive and negative syndrome scale (PANSS) for schizophrenia. Schizophr. Bull. 13: 261-276.

Kirkpatrick B. and Carpenter W. T. (1995): Drug development and the deficit syndrome of schizophrenia. Biolog. Psychiatry, 38: 277-278.

McGuire P. K., Silberweig D. A., Wright I., Murray R. M., Frackowiak R. S. J. and Frith C. D. (1996): The neural correlates of inner speech and auditory verbal imagery in schizophrenia: Relationship to auditory verbal hallucination. British Journal of Psychiatry, 169: 148-159.

McGuire P. K., Shah G. M. S. and Murray R. M. (1993): Increased blood flow in Broca's area during auditory hallucinations in schizophrenia. Lancet, 342: 703-706.

Neylan Th. C., van Kammen D. P., Kelley M. E., and Peters J. L. (1992): Sleep in schizophrenic patients on and off haloperidol therapy. Arch. Gener. Psychiatry, 49: 643-649.

Posner M., Early ., Reiman E., Pardo P., and Dhawan M. (1988): Asymmetries in hemispheric control of attention in schizophrenia. Arch. Gener. Psychiatry, 45: 814-821.

Rechtschaffen A. and Kales A. (Eds.) (1968): A manual of standardized terminology, techniques and scoring system for sleep stages of human subjects. Washington, DC: National Institute of Health, NIH publication 204.

Reynolds C. F. Ill, Kupfer D. J., Hauch P. R., Hoch C. C., Stack J. A., Berman S. R., Zimmer B. (1988): Reliable discrimination of elderly depressed and demented patients by electroencephalographic sleep data. Arch. Gener. Psychiatr. 45: 258-264.

Rotenberg V. S. (1984): Search activity in the context of psychosomatic disturbances, of brain monoamines and REM sleep function. Pavlov. J. Biol. Sciences, 19: 1-15.

Rotenberg V. S. (1988): Functional deficiency of REM sleep and its role in the pathogenesis of neurotic and psychosomatic disturbances. Pavlov. J. of Biolog. Sciences, 23: 1-3.

Rotenberg V. S. (1993): REM sleep and dreams as mechanism of search activity recovery. In: A. Moffitt, M. Kramer, R. Hoffmann (Eds.): Functions of dreaming, State University of New York Press, 261-292.

Rotenberg V. S. (1994): An integrative psychophysiological approach to brain hemisphere functions in schizophrenia. Neuroscience and Biobehavioral Reviews, 18: 487-495.

Rotenberg V. S. and Arshavsky V. V. (1979): Search activity and its impact on experimental and clinical pathology. Activit. Nervosa Superior, 21: 105-115.

Rotenberg V. S., Sirota P. and Elizur A. (1996): Psychoneuroimmunology: Searching for the main deteriorating psychobehavioral factor. Genetic, Social and General Psychology Monographs, 1996, 122: 329-346.

Tandon J.F.,Eiser A., de Quardo J. and Goodson J. (1992): Electroencephalographic sleep abnormalities in schizophre nia. Relationship to positive/negative symptoms and prior neuroleptic treatment. Arch. Gener. Psychiatr. 49: 185-194.

Tower R. and Flor-Henry P. (1989): Neuroleptics reverse attention asymmetries in schizophrenic patients. Biolog. Psychiatry 25: 852-860.