BRAIN ROBOT AUGMENTED INTERACTION
Our Brain Robot Augmented InteractioN (BRAIN) Lab is connecting robot to brain and creating innovations that help humans to lead their better life. To this end, our research aims at breakthroughs going well beyond the state of the art in Brain Robot Interaction.
March 1, 2023
CHUNGHO LEE A, JINUNG AN
EXPERT SYSTEMS WITH APPLICATIONS
LSTM-CNN MODEL OF DROWSINESS DETECTION FROM MULTIPLE CONSCIOUSNESS STATES ACQUIRED BY EEG
This study aimed to design a deep neural network for electroencephalography (EEG)-based drowsiness detection in multiple consciousness states, i.e., “awake,” “sleep,” and “drowsiness.” Few studies have seriously considered the optimal input vector size or labeling method in classifying multiple consciousness states, which may affect classification performance. To determine the optimal input vector length, i.e., window length, three neural network models (long short-term memory [LSTM], convolutional neural network [CNN], and combined LSTM and CNN) and four feature-based models were tested with six different levels of window length. The EEG dataset was acquired from 19 participants with randomly assigned auditory stimuli and button responses. The EEG data were labeled into three classes (awake, sleep, and drowsiness) based on the defined button response pattern corresponding to the stimuli. The results demonstrated that when the input vector size exceeded 8 sec, the performance of the neural network models dropped rapidly; however, when the window size was less than 8 sec, the performance change according to the window size was small. In contrast, the performance of feature-based models increased continuously as the window size increased. The LSTM model yielded the best accuracy (86%) for a 1 sec window length, and the LSTM-CNN model yielded the best kappa index (0.77) for a 4 sec window length. In addition, the proposed model was applied to the binary classification of normal consciousness (awake) and low consciousness (drowsiness and sleep) states to determine whether this model works appropriately in actual applications such as drowsiness detection in a driving environment. For binary classification, the LSTM-CNN model resulted in 0.95 F1 scores in 4000-ms. When a short input data (500 msec) is used, the LSTM-CNN model resulted in an average accuracy of 85.6% and a kappa index of 0.77 for the three-class classification problem and 0.94 F1 scores for the binary classification problem. In conclusion, we demonstrated that the proposed model could effectively detect drowsiness. Furthermore, a significant correlation was found between reaction time and drowsiness. However, using the reaction time as an index for labeling drowsiness was challenging because of the high false-negative ratio.
January 1u7, 2022
YEONGMI HA, SANG-HO LEE, DONG-HA LEE, YOUNG-HUN KANG, WOONJOO CHOI AND JINUNG AN
EFFECTIVENESS OF A MOBILE WELLNESS PROGRAM FOR NURSES WITH ROTATING SHIFTS DURING COVID-19 PANDEMIC: A PILOT CLUSTER-RANDOMIZED TRIAL
Nurses with rotating shifts, including night shifts, have suffered from low physical activity during the COVID-19 pandemic and lower sleep quality due to the disruption of their circadian rhythm. This study aimed to develop and examine the effectiveness of a mobile wellness program on daily steps, sleep quality, exercise self-efficacy, intrinsic motivation for exercise, self-rated fatigue, and wellness. A cluster randomized controlled trial design was used to examine the effectiveness of the mobile wellness program for nurses with rotating shifts. Sixty nurses from one university hospital participated and were allocated to an intervention group and a control group. The intervention group received a 12-week mobile wellness program to improve their physical activity and sleep quality, and the control group was only given a Fitbit to self-monitor their health behaviors. There were significant differences between the two groups in daily steps (p = 0.000), three components (subjective sleep quality, sleep disturbance, daytime dysfunction) of the PSQI, exercise self-efficacy, intrinsic motivation for exercise, and wellness. In conclusion, this study provides meaningful information that the mobile wellness program using Fitbit, online exercise using Zoom, online health coaching on a Korean mobile platform, and motivational text messages effectively promoted physical activity and sleep quality for nurses with rotating shifts during the COVID-19 pandemic.
Aug 8, 2020
SANG HYEON JIN, SEUNG HYUN LEE, SEUNG TAE YANG, JINUNG AN*
HEMISPHERIC ASYMMETRY IN HAND PREFERENCE OF RIGHT-HANDERS FOR PASSIVE VIBROTACTILE PERCEPTION: AN FNIRS STUDY
Hemispheric asymmetry in hand preference for passive cutaneous perception compared to active haptic perception is not well known. A functional near-infrared spectroscopy was used to evaluate the laterality of cortical facilitation when 31 normal right-handed participants were involved in 205 Hz passive vibrotactile cutaneous stimuli on their index fingers of preferred and less-preferred hand. Passive cutaneous perception resulted that preferred (right) hand stimulation was strongly leftward lateralized, whereas less-preferred (left) hand stimulation was less lateralized. This confirms that other manual haptic exploration studies described a higher hemispheric asymmetry in right-handers. Stronger cortical facilitation was found in the right primary somatosensory cortex (S1) and right somatosensory association area (SA) during left-hand stimulation but not right-hand stimulation. This finding suggests that the asymmetric activation in the S1 and SA for less-preferred (left) hand stimulation might contribute to considerably reinforce sensorimotor network just with passive vibrotactile cutaneous stimulation.
Jin, S.H., Lee, S.H., Yang, S.T. et al. Hemispheric asymmetry in hand preference of right-handers for passive vibrotactile perception: an fNIRS study. Sci Rep 10, 13423 (2020).
Oct 01, 2019
SANG HYEON JIN, SEUNG HYUN LEE, JINUNG AN*
THE DIFFERENCE IN CORTICAL ACTIVATION PATTERN FOR COMPLEX MOTOR SKILLS: A FUNCTIONAL NEAR- INFRARED SPECTROSCOPY STUDY
The human brain is lateralized to dominant or non-dominant hemispheres, and controlled through large-scale neural networks between correlated cortical regions. Recently, many neuroimaging studies have been conducted to examine the origin of brain lateralization, but this is still unclear. In this study, we examined the differences in brain activation in subjects according to dominant and non-dominant hands while using chopsticks. Fifteen healthy right-handed subjects were recruited to perform tasks which included transferring almonds using stainless steel chopsticks. Functional near-infrared spectroscopy (fNIRS) was used to acquire the hemodynamic response over the primary sensory-motor cortex (SM1), premotor area (PMC), supplementary motor area (SMA), and frontal cortex. We measured the concentrations of oxy-hemoglobin and deoxy-hemoglobin induced during the use of chopsticks with dominant and non-dominant hands. While using the dominant hand, brain activation was observed on the contralateral side. While using the non-dominant hand, brain activation was observed on the ipsilateral side as well as the contralateral side. These results demonstrate dominance and functional asymmetry of the cerebral hemisphere.
Jun 1, 2018
SEUNG HYUN LEE, SANG HYEON JIN, AND JINUNG AN*
DISTINCTION OF DIRECTIONAL COUPLING IN SENSORIMOTOR NETWORKS BETWEEN ACTIVE AND PASSIVE FINGER MOVEMENTS USING FNIRS
The purpose of this study is to investigate cerebral cortex activation during active movement and passive movement by using a functional near-infrared spectroscopy (fNIRS). Tasks were the flexion/extension of the right hand finger by active movement and passive movement. Oxy-hemoglobin concentration changes calculated from fNIRS and analyzed the activation and connectivity so as to understand dynamical brain relationship. The results demonstrated that the brain activation in passive movements is similar to motor execution. During active movement, the estimated causality patterns showed significant causality value from the supplementary motor area (SMA) to the primary motor cortex (M1). During the passive movement, the causality from the primary somatosensory cortex (S1) to the primary motor cortex (M1) was stronger than active movement. These results demonstrated that active and passive movements had a direct effect on the cerebral cortex but the stimulus pathway of active and passive movement is different. This study may contribute to better understanding how active and passive movements can be expressed into cortical activation by means of fNIRS.
Lee SH, Jin SH, An J. Distinction of directional coupling in sensorimotor networks between active and passive finger movements using fNIRS. Biomed Opt Express. 2018;9(6):2859-2870. Published 2018 May 31. doi:10.1364/BOE.9.002859