Working memory load recognition with deep learning time series classification.

Working memory load (WML) is one of the widely applied signals in the areas of human-machine interaction. The precise evaluation of the WML is crucial for this kind of application. This study aims to propose a deep learning (DL) time series classification (TSC) model for inter-subject WML decoding. We used fNIRS to record the hemodynamic signals of 27 participants during visual working memory tasks. Traditional machine learning and deep time series classification algorithms were respectively used for intra-subject and inter-subject WML decoding from the collected blood oxygen signals. The intra-subject classification accuracy of LDA and SVM were 94.6% and 79.1%. Our proposed TAResnet-BiLSTM model had the highest inter-subject WML decoding accuracy, reaching 92.4%. This study provides a new idea and method for the brain-computer interface application of fNIRS in real-time WML detection.

Reorganization of prefrontal network in stroke patients with dyskinesias: evidence from resting-state functional near-infrared spectroscopy.

Stroke usually causes multiple functional disability. To develop novel rehabilitation strategies, it is quite necessary to improve the understanding of post-stroke brain plasticity. Here, we use functional near-infrared spectroscopy to investigate the prefrontal cortex (PFC) network reorganization in stroke patients with dyskinesias. The PFC hemodynamic signals in the resting state from 16 stroke patients and 10 healthy subjects are collected and analyzed with the graph theory. The PFC networks for both groups show small-world attributes. The stroke patients have larger clustering coefficient and transitivity and smaller global efficiency and small-worldness than healthy subjects. Based on the selected network features, the established support vector machine model classifies the two groups of subjects with an accuracy rate of 88.5%. Besides, the clustering coefficient and local efficiency negatively correlate with patients' motor function. This study suggests that the PFC of stroke patients with dyskinesias undergoes specific network reorganization.

Age-related difference in muscle metabolism patterns during upper limb's encircling exercise: a near-infrared spectroscopy study.

Aging is usually accompanied by decrease in limb motor function and change in muscle metabolism patterns. However, few studies have investigated the aging effect on muscle hemodynamics of the upper extremity. This study aims to explore the aging effect on muscle metabolism patterns during upper limb's exercise. Twelve middle-aged and elderly subjects and 12 young subjects were recruited, and muscle oxygenation signals from these subjects' biceps brachii muscles were collected during active and passive upper limb's encircling exercise with near-infrared spectroscopy (NIRS). The old group showed stronger muscle hemodynamic metabolism than the young group. The multiscale fuzzy approximate entropy and multiscale transfer entropy analyses indicated higher complexity and stronger interlimb coupling of the muscle oxygenation signals for the old group. Based on the selected muscle metabolism features, the constructed support vector machine model showed a high accuracy rate for classifying the two groups of subjects: 91.6% for the passive mode and 87.5% for the active mode. Our results proved the specific muscle metabolism patterns in the upper limb's exercise for old subjects, promoting the understanding of the aging effect on muscle hemodynamics.