Using neuromorphic computing in prediction of GABA concentration - a pilot study.

Area of Science:

• Neuroscience
• Computer Science
• Biophysics

Background:

• Neuromorphic computing offers energy-efficient, real-time machine learning for biological signal processing.
• Accurate detection of Gamma-Aminobutyric Acid (GABA) concentration is crucial for understanding brain function.

Purpose of the Study:

• To investigate the feasibility of using neuromorphic computing for GABA concentration detection.
• To compare machine learning models for predicting GABA concentration from dielectric properties.

Main Methods:

• Dielectric relaxation spectroscopy was employed to gather permittivity data from various GABA solutions.
• Feedforward Neural Networks (FFNN) and Convolutional Neural Networks (CNN) were trained and compared for prediction accuracy.
• The CNN model was converted into Spiking Neural Networks (SNNs) for neuromorphic implementation.

Main Results:

• The study demonstrated the potential of neuromorphic computing for GABA concentration detection.
• Spiking Neural Networks (SNNs) exhibited promising energy efficiency and real-time processing capabilities.
• A user-friendly system integrating Tkinter facilitated seamless data transfer for flexibility and scalability.

Conclusions:

• Neuromorphic computing, particularly SNNs, presents a viable and efficient approach for real-time GABA concentration monitoring.
• This technology could advance brain-computer interfaces and neurological disorder diagnostics.
• The developed system offers a scalable and user-friendly platform for neuromorphic-based biosensing.