Spatiotemporal characteristics of cortical activities of REM sleep behavior disorder revealed by explainable machine learning using 3D convolutional neural network.

Isolated rapid eye movement sleep behavior disorder (iRBD) is a sleep disorder characterized by dream enactment behavior without any neurological disease and is frequently accompanied by cognitive dysfunction. The purpose of this study was to reveal the spatiotemporal characteristics of abnormal cortical activities underlying cognitive dysfunction in patients with iRBD based on an explainable machine learning approach. A convolutional neural network (CNN) was trained to discriminate the cortical activities of patients with iRBD and normal controls based on three-dimensional input data representing spatiotemporal cortical activities during an attention task. The input nodes critical for classification were determined to reveal the spatiotemporal characteristics of the cortical activities that were most relevant to cognitive impairment in iRBD. The trained classifiers showed high classification accuracy, while the identified critical input nodes were in line with preliminary knowledge of cortical dysfunction associated with iRBD in terms of both spatial location and temporal epoch for relevant cortical information processing for visuospatial attention tasks.

Characterization of attentional event-related potential from REM sleep behavior disorder patients based on explainable machine learning.

BACKGROUND AND OBJECTIVE: Idiopathic rapid eye movement sleep behavior disorder (iRBD) is a prodromal stage of neurodegeneration and is associated with cortical dysfunction. The purpose of this study was to investigate the spatiotemporal characteristics of cortical activities underlying impaired visuospatial attention in iRBD patients using an explainable machine-learning approach. METHODS: An algorithm based on a convolutional neural network (CNN) was devised to discriminate cortical current source activities of iRBD patients due to single-trial event-related potentials (ERPs), from those of normal controls. The ERPs from 16 iRBD patients and 19 age- and sex-matched normal controls were recorded while the subjects were performing visuospatial attentional task, and converted to two-dimensional images representing current source densities on flattened cortical surface. The CNN classifier was trained based on overall data, and then, a transfer learning approach was applied for the fine-tuning to each patient. RESULTS: The trained classifier yielded high classification accuracy. The critical features for the classification were determined by layer-wise relevance propagation, so that the spatiotemporal characteristics of cortical activities that were most relevant to cognitive impairment in iRBD were revealed. CONCLUSIONS: These results suggest that the recognized dysfunction in visuospatial attention of iRBD patients originates from neural activity impairment in relevant cortical regions and may contribute to the development of useful iRBD biomarkers based on neural activity.

Explainable Machine-Learning-Based Characterization of Abnormal Cortical Activities for Working Memory of Restless Legs Syndrome Patients.

Restless legs syndrome (RLS) is a sensorimotor disorder accompanied by a strong urge to move the legs and an unpleasant sensation in the legs, and is known to accompany prefrontal dysfunction. Here, we aimed to clarify the neural mechanism of working memory deficits associated with RLS using machine-learning-based analysis of single-trial neural activities. A convolutional neural network classifier was developed to discriminate the cortical activities between RLS patients and normal controls. A layer-wise relevance propagation was applied to the trained classifier in order to determine the critical nodes in the input layer for the output decision, i.e., the time/location of cortical activities discriminating RLS patients and normal controls during working memory tasks. Our method provided high classification accuracy (~94%) from single-trial event-related potentials, which are known to suffer from high inter-trial/inter-subject variation and low signal-to-noise ratio, after strict separation of training/test/validation data according to leave-one-subject-out cross-validation. The determined critical areas overlapped with the cortical substrates of working memory, and the neural activities in these areas were correlated with some significant clinical scores of RLS.

Spatiotemporal Characteristics of Event-Related Potentials Triggered by Unexpected Events during Simulated Driving and Influence of Vigilance.

We investigated the spatiotemporal characteristics of brain activity due to sudden events during monotonous driving and how it changes with vigilance level. Two types of sudden events, emergency stop and car drifting, were presented using driving simulator, and event-related potentials (ERPs) were measured. From the ERPs of both types of events, an early component representing sensory information processing and a late component were observed. The early component was expected to represent sensory information processing, which corresponded to visual and somatosensory/vestibular information processing for the sudden stop and lane departure tasks, respectively. The late components showed spatiotemporal characteristics of the well-known P300 component for both types of events. Common characteristic brain activities occurred in response to sudden events, regardless of the type. The modulation of brain activity due to the vigilance level also shared common characteristics between the two types. We expect that our results will contribute to the development of an effective means to assist drivers' reactions to ambulatory situations.

Emotional arousal due to video stimuli reduces local and inter-regional synchronization of oscillatory cortical activities in alpha- and beta-bands.

The purpose of current study is to reveal spatiotemporal features of oscillatory EEG activities in response to emotional arousal induced by emotional video stimuli, and to find the characteristics of cortical activities showing significant difference according to arousal levels. The EEGs recorded during watching affective video clips were transformed to cortical current density time-series, and then, cluster-based permutation test was applied to determine the spatiotemporal origins of alpha- and beta-band activities showing significant difference between high and low arousal levels. We found stronger desynchronization of alpha-band activities due to higher arousal in visual areas, which may be due to stronger activation for sensory information processing for the highly arousing video stimuli. In precentral and superior parietal regions, the stronger desynchronization in alpha-and low beta-bands was observed for the high arousal stimuli. This is expected to reflect enhanced mirror neuron system activities, which is involved in understanding the intention of other's action. Similar changes according to arousal level were found also in inter-regional phase synchronization in alpha- and beta-bands.