NapHack – Hack your naps for better productivity!
Date:
In these hectic times, power naps are a necessity to increase productivity with the crucial parts of effective napping include a relaxing sleep and waking up at the right sleep-state. To address this need, we are developing NapHack, an application that uses auditory neurofeedback to assist sleeping in a relaxed state of mind and monitoring sleep states in real-time to wake-up the user at a predetermined sleep-state. The python packages we use are muse-lsl, audiostream, numpy, mne, pickle, scipy, pandas, matplotlib, and pyedflib. We used muse-lsl to stream real-time EEG data (Fs=256Hz) from 4 electrodes of MUSE-S to a PC. The data is band-pass filtered from 1-40Hz (IIR filter) and z-scored. Power in delta (1-3Hz) and alpha (8-12Hz) bands is computed and their ratio is inversely mapped to the frequency of an audio-tone, used as neurofeedback. Higher alpha-to-delta ratio, indicating a relaxed state, would increasingly correspond to lower audio frequencies, enabling relaxed sleep. Additionally, preprocessed data is used to estimate spectral features (FFT), classify sleep states, and monitor them in real-time using a pre-trained SVM classifier. After the desired nap time and upon reaching a predetermined sleep state, an alarm is set-off to wake-up the user. The SVM classifier is trained to classify 6 sleep-states using EEG data acquired from publicly available CAP Sleep Database, and provides a test accuracy of 57% (Chance=16.7% ). Through a novel focus on optimizing naps, NapHack is paving the way for the future of napping – choose health, choose wellness and choose NapHack!