Treatment-resistant schizophrenia (TRS) impacts roughly one-third of patients who do not respond to adequate trials of dopamine (DA)-blocking antipsychotics. About half of these also fail clozapine, leading to especially poor outcomes. TRS often appears early—most cases are resistant from first episode—though a minority convert to resistance later, potentially via relapse-related mechanisms. Predictors include younger onset, longer untreated psychosis, prominent negative symptoms, cognitive impairment, poorer premorbid function, obstetric complications, neurological soft signs, and higher familial risk. Converging biology suggests TRS differs from treatment-responsive illness: normal striatal DA synthesis capacity with elevated anterior cingulate glutamate, greater cortical thinning/atrophy and dysconnectivity (notably prefrontal/frontotemporal), and hints of distinct genetic burden.

Clozapine remains the gold standard and should be initiated earlier, yet is underused and delayed, with “ultra-TRS” common. Pharmacologic augmentation shows limited benefit and more side effects, though ECT can aid short-term outcomes and repetitive trans-magnetic stimulation and transcranial direct stimulation may help specific symptoms. Given the likely glutamatergic/GABAergic and endocannabinoid involvement, novel targets are promising. Clinically, the field needs reliable biomarkers and simple predictive tools to identify DA-nonresponders early and triage to clozapine or alternatives, alongside standardized definitions and multicenter validations to stratify patients and personalize therapy.

Reference: Leung CC, Gadelrab R, Ntephe CU, McGuire PK, Demjaha A. Clinical Course, Neurobiology and Therapeutic Approaches to Treatment Resistant Schizophrenia. Toward an Integrated View. Front Psychiatry. 2019;10:601. doi: 10.3389/fpsyt.2019.00601.

Authors of this first large-scale scientometric map of negative-symptom research in schizophrenia analyzed 27,568 papers (322,349 citations; 59,905 authors) retrieved from Web of Science and visualized with Bibliometrix and CiteSpace. From 1980 to 2021, two long arcs dominate: (1) conceptualization/assessment—moving from the  Scale for the Assessment of Negative Symptoms/Positive and Negative Syndrome Scale era to finer, multi-domain models and distinctions between primary vs secondary negative symptoms; and (2) treatment—evolving from atypical antipsychotics and Clinical Antipsychotic Trials of Intervention Effectiveness-era trials to expansive evidence syntheses and long-acting injectables, alongside adjacent lines in neuroimaging, cognition, reward processing, inflammation/biomarkers, and early-psychosis cohorts.

Recent hotspots emphasize evidence synthesis for negative-symptom treatments, nonpharmacologic approaches (brain stimulation, cognitive remediation, exercise), computational psychiatry/machine learning, and ecological momentary assessment. Influential anchors include the Clinical Assessment Interview for Negative Symptoms (2013), a 2015 meta-analysis of treatments, and updated guidelines. Noted gaps include links with substance use, older adults, and health-resource impact. Limitations include citation/database biases and delayed recognition of novel work. The map guides clinicians, researchers, and funders to pivotal papers, collaborators, venues, and emerging themes for reviews and grants.

Reference: Sabe M, Chen C, Perez N, et al. Thirty years of research on negative symptoms of schizophrenia: A scientometric analysis of hotspots, bursts, and research trends. Neurosci Biobehav Rev. 2023;144:104979. doi: 10.1016/j.neubiorev.2022.104979.

Mental illnesses are common, multifactorial disorders that disrupt cognition, emotion, and behavior, reducing quality of life and life expectancy. High-burden conditions such as major depressive disorder and schizophrenia often begin early and vary widely across regions and sexes. For example, depressive disorders show a markedly higher Disability-Adjusted Life Year burden in women in several countries, while schizophrenia tends to be slightly higher in men. These patterns underscore the need for prevention, early diagnosis, and coordinated, interdisciplinary care using validated tools alongside clinical assessment frameworks.

Recent work highlights artificial intelligence’s growing role in earlier, more personalized detection: studies examine blood-based gene expression, neuroimaging, and real-world signals like motor activity. Building on this, the authors transform 24-hour wrist-actigraphy into images and train a two-dimensional convolutional neural network to classify control, depression, and schizophrenia using two public datasets. The model performed well, achieving about 84% accuracy on a held-out blind test (with lower recall for depression due to class imbalance), suggesting a feasible, noninvasive path for continuous monitoring—potentially via wearables. The authors recommend larger, more diverse cohorts, better label consistency, techniques to handle class imbalance, and rigorous clinical validation with attention to privacy and ethics.

Reference: Espino-Salinas CH, Luna-García H, Cepeda-Argüelles A, et al. Convolutional Neural Network for Depression and Schizophrenia Detection. Diagnostics (Basel). 2025;15(3):319. doi: 10.3390/diagnostics15030319.