Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Neuromorphic Heart Rate Monitors: Neural State Machines for Monotonic Change Detection

Alessio Carpegna, Chiara De Luca, Federico Emanuele Pozzi, Alessandro Savino, Stefano Di Carlo, Giacomo Indiveri, Elisa Donati

TL;DR

This work tackles the problem of detecting monotonic HR changes for continuous health monitoring by deploying a neural state machine (NSM) on a neuromorphic processor. The approach encodes HR-related states using EI-balanced populations within a WTA framework and gate-controlled transitions to enforce monotonic increases, achieving online processing with low power. Key findings show robust monotonic tracking on real ECG data with effective noise rejection and a total power budget around $90\mu W$, making it suitable for wearable wear-and-forget devices. The work promises practical impact for long-term health management, including dementia care, by enabling energy-efficient, always-on HR trend monitoring.

Abstract

Detecting monotonic changes in heart rate (HR) is crucial for early identification of cardiac conditions and health management. This is particularly important for dementia patients, where HR trends can signal stress or agitation. Developing wearable technologies that can perform always-on monitoring of HRs is essential to effectively detect slow changes over extended periods of time. However, designing compact electronic circuits that can monitor and process bio-signals continuously, and that can operate in a low-power regime to ensure long-lasting performance, is still an open challenge. Neuromorphic technology offers an energy-efficient solution for real-time health monitoring. We propose a neuromorphic implementation of a Neural State Machine (NSM) network to encode different health states and switch between them based on the input stimuli. Our focus is on detecting monotonic state switches in electrocardiogram data to identify progressive HR increases. This innovative approach promises significant advancements in continuous health monitoring and management.

Neuromorphic Heart Rate Monitors: Neural State Machines for Monotonic Change Detection

TL;DR

This work tackles the problem of detecting monotonic HR changes for continuous health monitoring by deploying a neural state machine (NSM) on a neuromorphic processor. The approach encodes HR-related states using EI-balanced populations within a WTA framework and gate-controlled transitions to enforce monotonic increases, achieving online processing with low power. Key findings show robust monotonic tracking on real ECG data with effective noise rejection and a total power budget around , making it suitable for wearable wear-and-forget devices. The work promises practical impact for long-term health management, including dementia care, by enabling energy-efficient, always-on HR trend monitoring.

Abstract

Detecting monotonic changes in heart rate (HR) is crucial for early identification of cardiac conditions and health management. This is particularly important for dementia patients, where HR trends can signal stress or agitation. Developing wearable technologies that can perform always-on monitoring of HRs is essential to effectively detect slow changes over extended periods of time. However, designing compact electronic circuits that can monitor and process bio-signals continuously, and that can operate in a low-power regime to ensure long-lasting performance, is still an open challenge. Neuromorphic technology offers an energy-efficient solution for real-time health monitoring. We propose a neuromorphic implementation of a Neural State Machine (NSM) network to encode different health states and switch between them based on the input stimuli. Our focus is on detecting monotonic state switches in electrocardiogram data to identify progressive HR increases. This innovative approach promises significant advancements in continuous health monitoring and management.
Paper Structure (10 sections, 4 figures, 1 table)

This paper contains 10 sections, 4 figures, 1 table.

Table of Contents

  1. Introduction
  2. Materials and Methods
  3. Neuromorphic Hardware
  4. Dataset and signal processing
  5. Network on chip
  6. Experimental Results
  7. WTA network: non-monotonic state transitions
  8. NSM network: monotonic state transitions
  9. Network dynamics on real ECG signal
  10. Conclusion

Figures (4)

  • Figure 1: Monotonic NSM. The input signal is filtered through four $4^{th}$ order Butterworth bandpass filters. Each filtered component is converted into spikes through a AdEx InF neuron (blue). The input spikes are fed into populations of AdEx InF neurons (blue), encoding different states of the network, interconnected in a WTA configuration via a common inhibitory population (red). States are connected to gating populations (yellow) to implement the monotonic computation. These are organized in a EI-balanced configuration with an inhibitory population (red) limiting the overall activity.
  • Figure 2: Network behavior when stimulated with 50Hz Poissonian sequences testing all the possible transitions: (a) non-monotonic WTA dynamics; (b) monotonic NSM response
  • Figure 3: Stable sustained activity of one of the EI-balanced primitives included in the networks
  • Figure 4: Response of the network when stimulated with a real ECG signal: (a) intense 10 minutes bike session; (b) 6 minutes and 40 seconds of walk