A High-Speed Stereo Camera System to Enable Non-Contact Strain Measurements at Extreme Magnifications and Temperatures

Abstract

A few notes can trigger recall of a favorite song. Can we crack the code by which such temporally unfolding memories are stored in the synaptic connections of the brain? Our team has developed a new mathematical approach to compute the weight matrices needed for a Recurrent Neural Network (RNN) to generate multiple specific Spatiotemporal Neural Activity Patterns (SNAPs), each activated by presentation of a short spatiotemporal trigger sequence. Here, we propose to download this mathematically defined weight matrix into biological neural tissue and trigger the replay of stored spatiotemporal activity patterns using theoretically defined trigger sequences. Using human fibroblasts that have been directly reprogrammed to form induced neurons (iNs) we will create neural monolayers on a next generation CMOS-based 4,096 channel stimulate-and-record system. Using patterned electrical stimulation sequences optimized to engage spike-timing-dependent plasticity (STDP), we will gradually adjust synaptic connectivity across the entire network to achieve the theoretically defined weight matrix. Changes in effective synaptic connectivity will be quantified throughout the training process using a stimulate-and-record protocol, and a gradient descent algorithm will be used to achieve the weight matrix needed to enable replay of stored SNAPs. In short, we will compose a neural symphony in the language of theoretical neuroscience, transcribe it into the synapses of a population of human IPSC-derived neurons, and trigger its replay across the neural population. This will be a first in neuroscience and will, as detailed below, open a whole new approach to understanding synaptic plasticity and neural computation.

Document Details

Document Type
DoD Grant Award
Publication Date
Feb 05, 2025
Source ID
FA95502410029

Entities

People

  • Ryan Berke

Organizations

  • Air Force Office of Scientific Research
  • United States Air Force
  • Utah State University

Tags

Fields of Study

  • Biology

Readers

  • Neural Network Machine Learning.
  • Neuroscience

Technology Areas

  • AI & ML
  • AI & ML - Neural Networks