Abstract:Today, more than 90% of inpatients hospitalized with Major Depression or Schizophrenia are treated with psychotropic drugs. Since none of the treatment options is causal, response rates are modest and the course of recovery is very heterogeneous. Genetic studies on the etiology and pathogenesis of major psychiatric disorders over the past decades have been largely unsuccessful. Likewise, genetic studies to predict response to psychopharmacological treatment have also not been particularly successful. In this project we have recruited 902 inpatients with ICD-10 diagnoses of schizophrenic ("F2 patients") or depressive disorders ("F3 patients"). The study assessed today's acute inpatient treatment regimens with up to 8 repeated measurements regarding the time course of recovery and adverse side effects. The genotyping included 100 candidate genes with genotypic patterns computed from 549 Single Nucleotide Polymorphisms (SNPs). To predict response to psychopharmacological treatment, we relied on a multidimensional approach to analyzing genetic diversity in combination with multilayer Neural Nets (NNs). Central to this new method were the "gene vectors" that (1) assessed the multidimensional genotypic patterns observed with genes; and (2) evaluated the correlations between genes. By means of these methods, we searched for combinations of multidimensional genotypic patterns that were characteristic of treatment responders while being rare among non-responders. The chosen method of approach provided a powerful technique to detail the complex structures of SNP data that are not detectable by conventional association methods. Molecular-genetic NNs enabled correct classification of 100% "non-responders", along with 94.7% correctly classified "responders" among the F2 patients, and 82.6% correctly classified "responders" among the F3 patients. The F2 and F3 classifiers were not disjoint but showed an overlap of 29.6% and 35.7% between the diagnostic groups, thus indicating that clinical diagnoses may not constitute etiologic entities. Our results suggested that patients may have an unspecific physical-genetic disposition that enables, facilitates, impedes or prevents recovery from major psychiatric disorders by setting various thresholds for exogenous triggers that initiate improvement ("recovery disposition"). Even though this disposition is not causally linked to recovery, it can nonetheless be clinically used in the sense of a "surrogate". Indeed, clinicians are also interested in reliable tools that can "do the job", despite the fact that etiology and pathogenesis of the treated disorders remain unknown.

Pathway-Phenotypes of Non-responders and Partial Responders to Treatment With Antipsychotics in Schizophrenia: A Machine Learning Study

COMT, 5-HTR2A, and SLC6A4 Mrna Expressions in First-Episode Antipsychotic-Naïve Schizophrenia and Association with Treatment Outcomes

Gut Microbial Biomarkers for the Treatment Response in First-Episode, Drug-Naïve Schizophrenia: a 24-Week Follow-Up Study

Neurocognition and NMDAR co-agonists pathways in individuals with treatment resistant first-episode psychosis: a 3-year follow-up longitudinal study

Predicting treatment resistance in schizophrenia patients: Machine learning highlights the role of early pathophysiologic features

Predictive Biomarkers for Antipsychotic Treatment Response in Early Phase of Schizophrenia: Multi-Omic Measures Linking Subcortical Covariant Network, Transcriptomic Signatures, and Peripheral Epigenetics

Stratification and prediction of remission in first-episode psychosis patients OPTIMISE

Incidence and factors associated of early non-response in first-treatment and drug-naïve patients with schizophrenia: a real-world study

A machine-learning framework for robust and reliable prediction of short- and long-term treatment response in initially antipsychotic-naïve schizophrenia patients based on multimodal neuropsychiatric data

High Mobility Group Protein 1 and Dickkopf-Related Protein 1 in Schizophrenia and Treatment-Resistant Schizophrenia: Associations with Interleukin-6, Symptom Domains, and Neurocognitive Impairments

The Neuroimmune and Neurotoxic Fingerprint of Major Neurocognitive Psychosis or Deficit Schizophrenia: a Supervised Machine Learning Study

Methylome-wide and meQTL analysis helps to distinguish treatment response from non-response and pathogenesis markers in schizophrenia

The novel pathway phenotype “major neurocognitive psychosis” is validated as a distinct class through the analysis of immune-linked neurotoxicity biomarkers and neurocognitive deficits

Neuromelanin-Sensitive MRI as Candidate Marker for Treatment Resistance in First-Episode Schizophrenia

Genetic determinants of antidepressant and antipsychotic drug response

Major Neurocognitive Psychosis: a Novel Schizophrenia Endophenotype Class That is Based on Machine Learning and Resembles Kraepelin's and Bleuler's Conceptions.

Neurodevelopmental predictors of treatment response in schizophrenia and bipolar disorder

Treatment-Resistant Schizophrenia: Genetic and Neuroimaging Correlates

Towards a New Classification of Stable Phase Schizophrenia into Major and Simple Neuro-Cognitive Psychosis: Results of Unsupervised Machine Learning Analysis

Predicting Treatment Response in Schizophrenia With Magnetic Resonance Imaging and Polygenic Risk Score

Development of a multivariate prediction model for antidepressant resistant depression using reward-related predictors