Search for: All records

Electrical signatures characteristic of complex neurological activity and neuropsychiatric disease are embedded in electroencephalography (EEG) signal data. To firmly establish new correlations between these brain electrical pulses and cognition, behavior, and disorders, researchers must achieve adequate statistical power to validate and mitigate uncertainties in their findings. This necessitates the usage of extensive studies involving large volumes of raw EEG data files from multiple subjects, data which must be preprocessed before conducting further analysis. While conventional processing and analysis of these raw data have been performed using isolated physical lab environments and stovepiped applications, there is a growing necessity for processing and analysis solutions that enable distributed processing of large data collections. This study presents a novel microservices approach as an alternative and complementary solution for retrieving and preprocessing EEG signal data. The approach leverages serverless technologies to deliver a highly scalable solution for processing massive amounts of EEG data. Deployed within a public cloud environment, we assess the efficacy of this method when employing various container orchestration configurations. This work demonstrates the capability for substantial enhancements in processing speeds, particularly when dealing with extensive EEG datasets. 

Neurological disabilities cause diverse health and mental challenges, impacting quality of life and imposing financial burdens on both the individuals diagnosed with these conditions and their caregivers. Abnormal brain activity, stemming from malfunctions in the human nervous system, characterizes neurological disorders. Therefore, the early identification of these abnormalities is crucial for devising suitable treatments and interventions aimed at promoting and sustaining quality of life. Electroencephalogram (EEG), a non-invasive method for monitoring brain activity, is frequently employed to detect abnormal brain activity in neurological and mental disorders. This study introduces an approach that extends the understanding and identification of neurological disabilities by integrating feature extraction, machine learning, and visual analysis based on EEG signals collected from individuals with neurological and mental disorders. The classification performance of four feature approaches—EEG frequency band, raw data, power spectral density, and wavelet transform—is assessed using machine learning techniques to evaluate their capability to differentiate neurological disabilities in short EEG segmentations (one second and two seconds). In detail, the classification analysis is conducted under two conditions: single-channel-based classification and region-based classification. While a clear demarcation between normal (healthy) and abnormal (neurological disabilities) EEG metrics may not be evident, their similarities and distinctions are observed through visualization, employing wavelet features. Notably, the frontal brain region (frontal lobe) emerges as a crucial area for distinguishing abnormalities among different brain regions. Also, the integration of wavelet features and visual analysis proves effective in identifying and understanding neurological disabilities. 

Electroencephalogram (EEG) research continues to rely heavily on data silos used in isolated physical lab environments. However, as a part of the digital transformation, the EEG community has begun its exploration of the public cloud to determine how it can be best utilized to increase collaboration and accelerate research outcomes. The growing number of online repositories for data and tools has provided additional computational resources but the process of downloading data and software along with the installation and configuration requirements is cumbersome and prone to error. To break away from this research paradigm, we present a novel application of cloud technologies to provide reusable EEG data acquisition and preprocessing software as a service (SaaS) that eliminates data and software downloading prerequisites. We utilize the Amazon Web Services (AWS) cloud platform and serverless technologies to create a distributed, highly scalable and extensible solution for EEG signal data preprocessing that is more conducive to effective collaboration and data reproducibility with the potential to expedite neurotechnology breakthroughs.