Psychiatry research · 2026
Classification of functional near-infrared spectroscopy (fNIRS) signals in schizophrenia and bipolar disorder using deep learning methods.
СтатьяНазвание и аннотация: Перевод пока не подготовлен
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Mou Lan, Gong Binbin, Tan Qian, Wang Hong, Zhang Xiaomei, Cui Ruifang, Wang Zefeng, Shen Zhongxia
Аннотация
BACKGROUND: Schizophrenia (SCZ) and Bipolar Disorder (BD) are prevalent mental disorders. This study uses functional near-infrared spectroscopy (fNIRS) combined with a verbal fluency task (VFT) to examine prefrontal function in SCZ and BD patients, differentiating them via hemodynamic change analysis. It also applies deep learning and interpretability methods to assess fNIRS' reliability as a clinical diagnostic tool and potential for differential diagnosis during cognitive tasks.
METHODS: A total of 50 SCZ, 67 BD patients and 52 healthy controls (HC) were enrolled. All underwent fNIRS to monitor prefrontal oxyhemoglobin (Oxy-Hb) concentration changes during VFT. Kruskal-Wallis test was used to compare Oxy-Hb changes across groups, with post-hoc attribution analysis to identify key fNIRS channels for classification.
RESULTS: Relative to HC, SCZ and BD patients exhibited markedly reduced prefrontal cortex activity during VFT. SCZ patients had lower activation levels than BD patients in the rFPC, lFPC, lDLPFC and lOFC (all P < 0.05). The three-category classification accuracy for the whole brain was 0.933 ± 0.044, with a mean AUC of 0.988 ± 0.017. Attribution analysis identified channel 17 as the top contributor to SCZ-BD differentiation.
CONCLUSION: The lOFC is expected to be the most critical neuroimaging biomarker for differentiating these two diseases. fNIRS has application value as an auxiliary diagnostic tool for mental diseases in clinical practice, and its diagnostic efficacy and clinical transformation potential are expected to be further enhanced when combined with deep learning technology.
METHODS: A total of 50 SCZ, 67 BD patients and 52 healthy controls (HC) were enrolled. All underwent fNIRS to monitor prefrontal oxyhemoglobin (Oxy-Hb) concentration changes during VFT. Kruskal-Wallis test was used to compare Oxy-Hb changes across groups, with post-hoc attribution analysis to identify key fNIRS channels for classification.
RESULTS: Relative to HC, SCZ and BD patients exhibited markedly reduced prefrontal cortex activity during VFT. SCZ patients had lower activation levels than BD patients in the rFPC, lFPC, lDLPFC and lOFC (all P < 0.05). The three-category classification accuracy for the whole brain was 0.933 ± 0.044, with a mean AUC of 0.988 ± 0.017. Attribution analysis identified channel 17 as the top contributor to SCZ-BD differentiation.
CONCLUSION: The lOFC is expected to be the most critical neuroimaging biomarker for differentiating these two diseases. fNIRS has application value as an auxiliary diagnostic tool for mental diseases in clinical practice, and its diagnostic efficacy and clinical transformation potential are expected to be further enhanced when combined with deep learning technology.