Skip to main content
2024 BMES Annual Meeting Main banner
Icon Legend
This session is not in your schedule.
This session is in your schedule. Click again to remove it.
Back

Device Technologies and Biomedical Robotics

Device Technologies and Biomedical Robotics - Poster Session C

Poster I20 - MyA: A Multi-Biometric Vest for People Living with Autism, Angelman Syndrome, and other Neurodevelopmental Disorders

Friday, October 25, 2024
10:00 AM – 11:00 AM EST
Location: Exhibit Hall E, F & G

Introduction:  Angelman syndrome (AS) is a neurodevelopmental disorder that causes limited speech, developmental delays, intellectual disability, seizures, and sensory sensitivities. There is no cure for AS; however, clinical trials testing interventions and therapies aimed at managing AS symptoms are underway. A major barrier to this research is a lack of suitable outcome measures for AS patients. Sensory sensitivities make common wearable monitors such as actigraphy watches and electroencephalography (EEG) caps intolerable to AS patients. This prevents the use of wearable devices to measure outcomes in many studies, and results in data loss when studies do attempt to use them. To remedy this, we have developed MyA, a physiological monitoring vest tailored to the unique sensory needs of AS patients (Figure 1). We specifically worked to refine MyA’s ability to capture high-quality electrocardiogram (ECG) signals using a non-invasive, textile-based, dry electrode configuration. AS patients are prone to frequent seizures, which can often lead to subtle changes in cardiac rhythm. Continuous ECG monitoring through MyA, can help clinicians correlate cardiac anomalies with seizure activity, improving care plans for AS patients. In addition to physiological signals, clinical psychologists can recognize changes in voice and speech patterns to study underlying behavioral and emotional states of AS patients leading to more holistic care strategies. We worked to integrate an audio recording device into MyA capable of collecting the high-definition audio recordings needed to analyze speech patterns in AS patients, helping clinical teams develop personalized plans for speech therapy and communication aids.

Materials and

Methods:  To maximize MyA’s wearability for AS patients, the ECG interface employs dry, textile electrodes. This makes acquiring high-quality signals more difficult than with standard wet, sticky electrode systems, particularly during periods of motion. We therefore tested a range of circuit configurations and filters to optimize our signal quality. The ECG module was comprised of the Analog Devices AD8232 single lead ECG analog front-end, an nRF52840 microcontroller, and an SD card slot, mounted on a custom-designed printed circuit board (PCB). Three PCBs with different AD8232 circuit configurations were designed in Autodesk EAGLE and manufactured by PCBWay. Each configuration was tested in the MyA vest across a range of movements including walking and waving arms. Data was filtered using MATLAB with high pass filters ranging from 10-60 Hz cutoff frequency.

For audio collection, the Teensy 4.1 microcontroller (PJRC) was tested with the Teensy Audio Adapter Board (SGTL5000 low-power audio stereo codec). A 20-year old female participant was positioned 20 cm from the microphone and recorded 15-20 seconds of audio while reading a generic sentence aloud. The audio data was sampled at 44.1 kHz, 16-bits, gain settings of 4 dB, 8 dB, 16 dB, and 32 dB, and was stored on the onboard SD card for offline analysis. The recorded audio data was processed and filtered offline using Audacity’s noise reduction feature (12 dB reduction) and then a spectrogram was generated in Praat along with pitch and intensity.

Results, Conclusions, and Discussions: The three AD8232 configurations we explored were the “Heart Rate Measurement Next to the Heart,” (Fig. 2) “Cardiac Monitor,” (Fig. 3) and “Exercise Application” (Fig. 4) configurations from its datasheet. The Exercise Application proved to be the most suitable, as it is specifically designed to minimize motion artifacts, which allows signal integrity to be maintained even during intense physical movements. This feature is vital in wearable devices, especially for AS patients who may exhibit hyperactivity. Suitable R-waves can be seen in the filtered data, making the Exercise Application configuration sufficient for determining heart rate, heart rate variability, and other derivative metrics. Heart rates computed during these trials ranged from 96-132 beats per minute (BPM) which is within expected rates during moderate activities such was walking. However, even with the built-in filtering, motion artifacts remained during intense periods of movement, particularly involving the arms and torso. Therefore, extracting finer details from the ECG such as ST elevation may not be possible. With these results, MyA has been shown to be capable of acquiring suitable R-waves for extracting heart rate across various states of activity, suggesting the device would be suitable for monitoring in the AS population.

For integrating audio, the Teensy 4.1 audio shield produced favorable results. Recorded audio was of sharp quality with only mild amounts of static and other unwanted background noise as depicted by the spectrogram and raw audio waveform shown in Figure 5. The median pitch computed by Praat was 231 Hz which is within the expected range for an adult female (165-255 Hz). Upon experimenting with different gain settings, we found that lower gain settings worked best to avoid clipping the audio. Consequently, the Teensy 4.1 microcontroller proved to be the optimal choice for the MyA integrated audio module, delivering superior performance in recording and analyzing speech data, making it an effective tool for both clinical and research analysis.

Acknowledgements (Optional):