Predicting Attentional Vulnerability to Sleep Deprivation: A Multivariate Pattern Analysis of DTI Data

• DOI: 10.2147/nss.s345328
• Publication Date: 2022-04-22T09:50:08Z
• Authors: Chen Wang; Peng Fang; Ya Li; Lin Wu; Tian Hu; Qi Yang; Aiping Han; Yingjuan Chang; Xing Tang; Xiuhua Lv; Ziliang Xu; Yongqiang Xu; Leilei Li; Minwen Zheng; Yuanqiang Zhu
• Keywords: Psychiatry; Neurophysiology and neuropsychology; QP351-495; vulnerability; RC435-571; psychomotor vigilance task; diffusion tensor imaging; 16. Peace & justice; sleep deprivation; 3. Good health; 03 medical and health sciences; machine learning; 0302 clinical medicine; Nature and Science of Sleep; Original Research

Large individual differences exist in sleep deprivation (SD) induced sustained attention deterioration. Several brain imaging studies have suggested that the activities within frontal-parietal network, cortico-thalamic connections, and inter-hemispheric connectivity might underlie neural correlates of vulnerability/resistance to SD. However, those traditional approaches are based on average estimates at group level. Currently, a neuroimaging marker can reliably predict this vulnerability level is lacking.Efficient transfer information relies integrity white matter (WM) tracts human brain, we therefore applied machine learning approach investigate whether WM diffusion metrics Forty-nine participants completed psychomotor vigilance task (PVT) both after resting wakefulness (RW) 24 h (SD). The number PVT lapse (reaction time > 500 ms) was calculated for RW condition SD were categorized as vulnerable (24 participants) or resistant (25 according change lapses between two conditions. Diffusion tensor acquired extract four multitype features regional level: fractional anisotropy, mean diffusivity, axial radial diffusivity. A linear support vector (LSVM) using leave-one-out cross-validation (LOOCV) performed assess discriminative power SD-vulnerable SD-resistant participants.LSVM analysis achieved correct classification rate 83.67% (sensitivity: 87.50%; specificity: 80.00%; area under receiver operating characteristic curve: 0.85) differentiating from participants. fiber contributed most model primarily commissural pathways (superior longitudinal fasciculus), projection (posterior corona radiata, anterior limb internal capsule) association (body genu corpus callosum). Furthermore, found significantly negative correlation changes LSVM decision value.These findings suggest fibers connecting (1) regions (2) thalamus prefrontal cortex, (3) left right hemispheres accuracy.