To present basic principles and requirements of a network consisting of research institutions and routine care facilities, which aims at the improvement of treatment and care in schizophrenia.

The concept, structure and the management of such a network will be exemplified by the German Research Network On Schizophrenia, which is funded by the German Ministry of Education and Research (BMBF).

The experiences so far make very clear, that it requires carefully tuned projects, an efficient and well-financed network management, acceptance and dissimination of the network idea within (and outside) the network as well as collateral political measures to improve the research environment to incorporate single research projects and single institutions, researchers or clinicians into a network.

These general conditions fulfilled, network research is a clever strategy to bundle competence (horizontally and vertically) and to improve treatment and care of psychiatric patients.

Disturbances of the oculomotor system are promising endophenotypes for schizophrenia. In two separate studies, we examined antisaccade task performance, a measure of inhibitory control, in first degree relatives of schizophrenic patients (genetic risk without manifest disorder) and in clinical high risk subjects with symptoms suggestive of a prodromal phase of schizophrenia.

In the first study, 41 parents of schizophrenia patients and 22 controls were tested with with a prosaccade task and an antisaccade task. Parents were grouped into more likely, less likely, and indeterminate risk carriers. The second study involved 160 subjects clinically at risk for schizophrenia, 32 first episode schizophrenic patients, and 76 healthy controls.

In study 1 we found an increase of antisaccade latencies and error rates in parents of schizophrenics which varied with inferred genetic load, more likely gene carriers performing worst. In study 2, antisaccade performance varied with symptom load: subjects at risk with basic symptoms only were unimpaired, while at-risk subjects who had experienced brief psychotic episodes (BLIPS) showed deficits similar to first episode patients.

Reduced inhibitory control of oculomotor performance is associated with genetic loading for schizophrenia, and also with symptoms placing subjects at imminent risk of psychosis.

Impairments in affect recognition are well known in schizophrenia. Such impairments are known to be a trait-like characteristic in schizophrenia mostly unaffected by traditional treatment. Moreover they seem to play a crucial role in patients' poor social functioning. The present study should contribute to the still open question of treatment options for these impairments.

A special Training of Affect Recognition (TAR) was evaluated using a pre-post-control group design with three groups of about n=25 partly remitted schizophrenia patients each. To control for nonspecific effects of implicit cognitive training, TAR was compared with a Cognitive Remediation Training (CRT) aiming at improvement of basic neurocognitive functioning. To control for nonspecific effects the two active training groups were compared with a control group without additional training (CG).

Patients under TAR showed an improvement in facial affect recognition, with recognition performance after training approaching the level of healthy controls from former studies. Patients under CRT and those without training (CG) did not show improvements in affect recognition, though patients under CRT improved in some memory functions.

Improvements in disturbed facial affect recognition in schizophrenia patients is not obtainable with a traditional cognitive remediation program like CRT, but needs a functional specific training like the newly developed TAR.

Psychosis is preceded by cognitive and physiological alterations. This may be useful in the risk assessment in subjects with putatively prodromal symptoms, and could contribute to better understand the temporal unfolding of the disease.

The early recognition and intervention program of the German Research Network on schizophrenia defines early and late prodromal stages according to psychopathological criteria. For concurrent and prospective validation of these risk stages, subjects undergo neurocognitive, electrophysiological and oculomotor assessments of putative vulnerability markers. About 125 early prodromal subjects (defined by the presence of basic symptoms, Klosterkoetter et al. 2001), and 90 late prodromal subjects (defined by attenuated positive symptoms or by brief occurrences of psychotic symptoms) have been assessed at inclusion.

As compared to psychiatrically healthy matched controls, late prodromals have significantly inferior verbal memory, verbal fluency, visual motor skills, and working memory. Impairments are qualitatively similar, but less pronounced in subjects in an early prodromal stage, with deficits of immediate verbal memory, verbal fluency and visuomotor performance being significant. Both groups show reduced auditory startle prepulse inhibition. Impairments are not correlated with depression and general distress scores, and are also largely independent of prodromal and attenuated positive symptoms. In early prodromals, global cognitive performance is related to the occurrence of psychotic symptoms during follow-up. Auditory P 300 is reduced in both prodromal groups, and predicts transitions to psychosis.

Neurocognitive and neurophysiological assessments validate and improve psychopathological risk assessment, and allow to disentangle stable vulnerability markers from indicators of imminent risk.

Funded by the German Federal Ministry for Education and Research BMBF (grant 01 GI 9934).