An interpretable TimeMIL framework for fNIRS: differential diagnosis between schizophrenia and bipolar disorder - Takeaways - MDSpire

An interpretable TimeMIL framework for fNIRS: differential diagnosis between schizophrenia and bipolar disorder

  • By

  • Zefeng Wang

  • Binbin Gong

  • Lan Mou

  • Qian Tan

  • Xinhua Shen

  • Ruifang Cui

  • June 10, 2026

  • 0 min

Share

  • 1

    The TimeMIL model effectively distinguishes schizophrenia from bipolar disorder using fNIRS data during verbal fluency tasks.

  • 2

    TimeMIL achieved an accuracy of 0.928 and a macro-averaged AUC of 0.984, outperforming traditional deep learning models.

  • 3

    The study highlights significant differences in the orbitofrontal cortex among healthy controls, schizophrenia, and bipolar disorder.

  • 4

    An interpretability framework was developed to enhance understanding of model predictions and neurophysiological underpinnings.

  • 5

    This research paves the way for objective neuroimaging tools for early screening and differential diagnosis of psychiatric disorders.

Original Source(s)

An interpretable TimeMIL framework for fNIRS: differential diagnosis between schizophrenia and bipolar disorder

  • by Zefeng Wang, Binbin Gong, Lan Mou, Qian Tan, Xinhua Shen, Ruifang Cui

  • June 10, 2026

Related Content

Gender-Diverse Youths Report More Psychotic Symptoms

Longitudinal cohort data linked bullying and persistently unsupportive state gender-identity policies with worsening psychotic-like experiences among gender-diverse youths.

Stewardship Cuts Steroid Prescribing

Behaviorally informed peer comparison feedback was associated with a gradual decline in glucocorticoid prescribing

Baseline working memory was associated with improvement in psychological quality of life in patients with persistent depressive symptoms: a prospective observational study