A low-cost and open-hardware portable 3-electrode sleep monitoring device

To continue sleep research activities during the lockdown resulting from the COVID-19 pandemic, experiments that were previously conducted in laboratories were shifted to the homes of volunteers. Furthermore, for extensive data collection, it is necessary to use a large number of portable devices. Hence, to achieve these objectives, we developed a low-cost and open-source portable monitor (PM) device capable of acquiring electroencephalographic (EEG) signals using the popular ESP32 microcontroller. The device operates based on instrumentation amplifiers. It also has a connectivity microcontroller with Wi-Fi and Bluetooth that can be used to stream EEG signals. This portable single-channel 3-electrode EEG device allowed us to record short naps and score different sleep stages, such as wakefulness, non rapid eye movement sleep (NREM), stage 1 (S1), stage 2 (S2), stage 3 (S3) and stage 4 (S4). We validated the device by comparing the obtained signals to those generated by a research-grade counterpart. The results showed a high level of accurate similarity between both devices, demonstrating the feasibility of using this approach for extensive and low-cost data collection of EEG sleep recordings.


Figure 1 :
Figure 1: Bland-Altman plots of EEG power NREM sleep recorded at Cz. Mean EEG Alpha power from the BrainAmp DC and Baby Blue (y-axis) plotted against the difference in mean power (y-axis).

Figure 2 :
Figure 2: Bland-Altman plots of EEG power S2 sleep recorded at Cz. Mean EEG Alpha power from the BrainAmp DC and Baby Blue (y-axis) plotted against the difference in mean power (y-axis).

Figure 3 :
Figure 3: Bland-Altman plots of EEG power S3 sleep recorded at Cz. Mean EEG Alpha power from the BrainAmp DC and Baby Blue (y-axis) plotted against the difference in mean power (y-axis).

Figure 4 :
Figure 4: Bland-Altman plots of EEG power S4 sleep recorded at Cz. Mean EEG Alpha power from the BrainAmp DC and Baby Blue (y-axis) plotted against the difference in mean power (y-axis).

Figure 5 :
Figure 5: Bland-Altman plots of EEG power SWS recorded at Cz. Mean EEG Alpha power from the BrainAmp DC and Baby Blue (y-axis) plotted against the difference in mean power (y-axis).

Figure 6 :
Figure 6: Bland-Altman plots of EEG power NREM sleep recorded at Cz. Mean EEG Beta power from the BrainAmp DC and Baby Blue (y-axis) plotted against the difference in mean power (y-axis).

Figure 7 :
Figure 7: Bland-Altman plots of EEG power S2 sleep recorded at Cz. Mean EEG Beta power from the BrainAmp DC and Baby Blue (y-axis) plotted against the difference in mean power (y-axis).

Figure 8 :
Figure 8: Bland-Altman plots of EEG power S3 sleep recorded at Cz. Mean EEG Beta power from the BrainAmp DC and Baby Blue (y-axis) plotted against the difference in mean power (y-axis).

Figure 9 :
Figure 9: Bland-Altman plots of EEG power S4 sleep recorded at Cz. Mean EEG Beta power from the BrainAmp DC and Baby Blue (y-axis) plotted against the difference in mean power (y-axis).

Figure 10 :
Figure 10: Bland-Altman plots of EEG power SWS recorded at Cz. Mean EEG Beta power from the BrainAmp DC and Baby Blue (y-axis) plotted against the difference in mean power (y-axis).

Figure 11 :
Figure 11: Bland-Altman plots of EEG power NREM sleep recorded at Cz. Mean EEG Delta power from the BrainAmp DC and Baby Blue (y-axis) plotted against the difference in mean power (y-axis).

Figure 12 :
Figure 12: Bland-Altman plots of EEG power S2 sleep recorded at Cz. Mean EEG Delta power from the BrainAmp DC and Baby Blue (y-axis) plotted against the difference in mean power (y-axis).

Figure 13 :
Figure 13: Bland-Altman plots of EEG power S3 sleep recorded at Cz. Mean EEG Delta power from the BrainAmp DC and Baby Blue (y-axis) plotted against the difference in mean power (y-axis).

Figure 14 :
Figure 14: Bland-Altman plots of EEG power S4 sleep recorded at Cz. Mean EEG Delta power from the BrainAmp DC and Baby Blue (y-axis) plotted against the difference in mean power (y-axis).

Figure 15 :
Figure 15: Bland-Altman plots of EEG power SWS recorded at Cz. Mean EEG Delta power from the BrainAmp DC and Baby Blue (y-axis) plotted against the difference in mean power (y-axis).

Figure 16 :
Figure 16: Bland-Altman plots of EEG power NREM sleep recorded at Cz. Mean EEG Fast Spindle power from the BrainAmp DC and Baby Blue (y-axis) plotted against the difference in mean power (y-axis).

Figure 17 :
Figure 17: Bland-Altman plots of EEG power S2 sleep recorded at Cz. Mean EEG Fast Spindle power from the BrainAmp DC and Baby Blue (y-axis) plotted against the difference in mean power (y-axis).

Figure 18 :
Figure 18: Bland-Altman plots of EEG power S3 sleep recorded at Cz. Mean EEG Fast Spindle power from the BrainAmp DC and Baby Blue (y-axis) plotted against the difference in mean power (y-axis).

Figure 19 :
Figure 19: Bland-Altman plots of EEG power S4 sleep recorded at Cz. Mean EEG Fast Spindle power from the BrainAmp DC and Baby Blue (y-axis) plotted against the difference in mean power (y-axis).

Figure 20 :
Figure 20: Bland-Altman plots of EEG power SWS recorded at Cz. Mean EEG Fast Spindle power from the BrainAmp DC and Baby Blue (y-axis) plotted against the difference in mean power (y-axis).

Figure 21 :
Figure 21: Bland-Altman plots of EEG power NREM sleep recorded at Cz. Mean EEG Slow Oscillation power from the BrainAmp DC and Baby Blue (y-axis) plotted against the difference in mean power (y-axis).

Figure 22 :
Figure 22: Bland-Altman plots of EEG power S2 sleep recorded at Cz. Mean EEG Slow Oscillation power from the BrainAmp DC and Baby Blue (y-axis) plotted against the difference in mean power (y-axis).

Figure 23 :
Figure 23: Bland-Altman plots of EEG power S3 sleep recorded at Cz. Mean EEG Slow Oscillation power from the BrainAmp DC and Baby Blue (y-axis) plotted against the difference in mean power (y-axis).

Figure 24 :
Figure 24: Bland-Altman plots of EEG power S4 sleep recorded at Cz. Mean EEG Slow Oscillation power from the BrainAmp DC and Baby Blue (y-axis) plotted against the difference in mean power (y-axis).

Figure 25 :
Figure 25: Bland-Altman plots of EEG power SWS recorded at Cz. Mean EEG Slow Oscillation power from the BrainAmp DC and Baby Blue (y-axis) plotted against the difference in mean power (y-axis).

Figure 26 :
Figure 26: Bland-Altman plots of EEG power NREM sleep recorded at Cz. Mean EEG Slow Spindle power from the BrainAmp DC and Baby Blue (y-axis) plotted against the difference in mean power (y-axis).

Figure 27 :
Figure 27: Bland-Altman plots of EEG power S2 sleep recorded at Cz. Mean EEG Slow Spindle power from the BrainAmp DC and Baby Blue (y-axis) plotted against the difference in mean power (y-axis).

Figure 28 :
Figure 28: Bland-Altman plots of EEG power S2 sleep recorded at Cz. Mean EEG Slow Spindle power from the BrainAmp DC and Baby Blue (y-axis) plotted against the difference in mean power (y-axis).

Figure 29 :
Figure 29: Bland-Altman plots of EEG power S4 sleep recorded at Cz. Mean EEG Slow Spindle power from the BrainAmp DC and Baby Blue (y-axis) plotted against the difference in mean power (y-axis).

Figure 30 :
Figure 30: Bland-Altman plots of EEG power SWS recorded at Cz. Mean EEG Slow Spindle power from the BrainAmp DC and Baby Blue (y-axis) plotted against the difference in mean power (y-axis).

Figure 31 :
Figure 31: Bland-Altman plots of EEG power NREM sleep recorded at Cz. Mean EEG Theta power from the BrainAmp DC and Baby Blue (y-axis) plotted against the difference in mean power (y-axis).

Figure 32 :
Figure 32: Bland-Altman plots of EEG power S2 sleep recorded at Cz. Mean EEG Theta power from the BrainAmp DC and Baby Blue (y-axis) plotted against the difference in mean power (y-axis).

Figure 33 :
Figure 33: Bland-Altman plots of EEG power S3 sleep recorded at Cz. Mean EEG Theta power from the BrainAmp DC and Baby Blue (y-axis) plotted against the difference in mean power (y-axis).

Figure 34 :
Figure 34: Bland-Altman plots of EEG power s4 sleep recorded at Cz. Mean EEG Theta power from the BrainAmp DC and Baby Blue (y-axis) plotted against the difference in mean power (y-axis).

Figure 35 :
Figure 35: Bland-Altman plots of EEG power SWS recorded at Cz. Mean EEG Theta power from the BrainAmp DC and Baby Blue (y-axis) plotted against the difference in mean power (y-axis).

Figure 36 :
Figure 36: Bland-Altman plots of EEG power NREM sleep recorded at Cz. Mean EEG Beta power from the BrainAmp DC and Baby Blue (y-axis) plotted against the difference in mean power (y-axis).Artifacts from BrainAmp DC signals were removed.

Figure 37 :
Figure 37: Bland-Altman plots of EEG power NREM sleep recorded at Cz. Mean EEG Slow Oscillation power from the BrainAmp DC and Baby Blue (y-axis) plotted against the difference in mean power (y-axis).Bandpass filter (0.48Hz -53.32Hz) was applied on BrainAmp DC signals.

Figure 38 :
Figure 38: Bland-Altman plots of EEG power NREM sleep recorded at Cz. Mean EEG Delta power from the BrainAmp DC and Baby Blue (y-axis) plotted against the difference in mean power (y-axis).Bandpass filter (0.48Hz -53.32Hz) was applied on BrainAmp DC signals.

Figure 39 :
Figure 39: Bland-Altman plots of EEG power NREM sleep recorded at Cz. Mean EEG Theta power from the BrainAmp DC and Baby Blue (y-axis) plotted against the difference in mean power (y-axis).Bandpass filter (0.48Hz -53.32Hz) was applied on BrainAmp DC signals.