Biomolecular Materials Emulate Short-Term Synaptic Plasticity for Signal Processing
Open Access
- Author:
- Almatar, Ahmed
- Area of Honors:
- Mechanical Engineering
- Degree:
- Bachelor of Science
- Document Type:
- Thesis
- Thesis Supervisors:
- Joseph Najem, Thesis Supervisor
Jean-Michel Mongeau, Thesis Honors Advisor - Keywords:
- biomolecular memristor
alamethicin
memristor
ion channel
biomembrane
neuromorphic computing
lipid bilayer
neuromorphic
DIB
smart materiels
biomolecular - Abstract:
- The amount of digital data we are producing is increasing at rapid rates and might soon exceed our current capacity to process it, mostly due to energy limitations. Therefore, to process this vast amount of data, optimizing computing per unit energy is key. Currently, neuromorphic computing, a model for computing that borrows key computational aspects of the human brain, is a leading solution to optimizing computing. However, despite major progress, solid-state neuromorphic computing hardware bears little resemblance to biological neurons and synapses, and thus, is still lagging in performance and energy efficiency compared to the brain. This study investigates a different approach to neuromorphic systems by using biomolecular memory elements as opposed to silicon-based elements. Replacing silicon with biomembranes to form memristors enables the emulation of short-term synaptic plasticity—a signal filtering property of biological synapses. The performance of the biomembrane has been tested experimentally using a DC signal as the input and a solid-state neuron circuit as the load. The biomembrane was determined to be exhibiting short-term synaptic plasticity by controlling the firing rate of the neuron. The results suggest that our biomolecular memristor is capable of performing basic signal processing tasks, namely, high-pass filtering. However, the behavior of networks of biomembranes remains unknown and is of great interest.