Suvam Dey

Slow conscious breathing is an integral aspect of many therapeutic techniques due to its relaxing effects. While the neuroscience of conscious breathing has been explored widely, the exact neural mechanisms linking slow breathing to its therapeutic effects are still debated. This work utilizes electroencephalography (EEG) to identify functional connections specific to the inhalation and exhalation phases of slow deep breathing at 2 cycles per minute. EEG data were collected from 20 healthy participants using the 61-channel eego™ mylab system from ANT Neuro. Functional connectivity (FC) for all possible pairs of EEG time series data was estimated using the phase slope index, for 7 EEG bands. Further, feature selection and classification were performed to identify functional connections that could effectively distinguish the inhale from the exhale phase of the respiratory cycle. The best accuracy of 99.08% was obtained when 340 low gamma-band functional connections were fed as input to a support vector machine with radial basis function kernel. Furthermore, the inter- and intra-cortical distribution of these functional connections was explored based on the topographical grouping of EEG electrodes. It was observed that most of the statistically significant connections were within central or between central and parieto-occipital regions of the brain.

In recent years, significant advances in biometrics have essentially been driven by machine learning (ML) and deep learning (DL) progress. Numerous human identification applications are currently available using physical traits such as fingerprints, face, and voice. With the development of Internet of Things (IoT) sensors and the availability of a variety of ML algorithms, there has been increased research interest in subject identification (SI) based on behavioral cues. For example, several research works have been published on SI based on gait analysis. Sensors like accelerometers (ACC), gyroscopes, (GYR), and magnetometers (MAG) were used to collect data during limited activities such as walking. We believe that using data for one activity is not sufficient to adequately capture behavioral cues for the purpose of SI. Considering other cues such as gestures or head shaking and using a variety of sensors located on different parts of the human body are essential to developing a scenario that includes expressive human activities. We designed a specific scenario that included several activities, such as walking, giving a talk, chatting while sitting, and climbing stairs, using five inertial measurement units (IMU) located on various parts of the human body. Several ML algorithms, namely Linear Discriminant Analysis (LDA), K-Nearest Neighbours (KNN), Random Forest (RF), and XGBoost (XGB) were used. Our results show that SI yields beyond 99% accuracy for most activities. Furthermore, we succeeded in implementing a real-time IoT system for SI based on our best offline results. We achieved 98.04% accuracy within 0.06 ms of processing time (PT).