Related Experiment Video
Updated: Jun 11, 2026
08:48
Optical Recording of Suprathreshold Neural Activity with Single-cell and Single-spike Resolution
Published on: September 5, 2012
Recording from two neurons: second-order stimulus reconstruction from spike trains and population coding
N M Fernandes1, B D L Pinto, L O B Almeida
1Instituto de Física de São Carlos, Universidade de São Paulo, São Carlos, Brazil. nelson@ifsc.usp.br
Neural Computation
|July 9, 2010
Summary
We reconstruct visual stimuli from fly H1 neuron spike trains using a Volterra series. Our method efficiently handles complex data, improving rotational stimulus reconstruction accuracy.
Area of Science:
- Neuroscience
- Computational Neuroscience
- Systems Neuroscience
Background:
- Reconstructing visual stimuli from neural activity is crucial for understanding sensory processing.
- Spike train analysis often employs models like the Volterra series, but higher-order terms can be computationally intensive.
- The H1 neurons in flies provide a well-studied model system for visual processing.
Purpose of the Study:
- To develop an efficient method for reconstructing visual stimuli from spiking neuron data using a second-order Volterra series.
- To investigate the contribution of second-order kernels to stimulus reconstruction in fly H1 neurons.
- To present a computational scheme applicable to large neural populations.
Main Methods:
- Utilized a second-order Volterra series to model the relationship between spike trains and visual stimuli.
- Employed a novel approach using basis functions to avoid large matrix computations and inversions.
- Approximated spike train four-point functions using two-point functions, analogous to Gaussian processes.
- Tested the method on simultaneous recordings from two H1 neurons in the fly Chrysomya megacephala responding to rotational and translational stimuli.
Main Results:
- The approximation of spike train functions did not compromise reconstruction quality.
- Second-order kernels contributed minimally (around 5% mean squared error) to overall stimulus reconstruction.
- However, second-order kernels significantly improved reconstruction (up to 100%) for rotational stimuli at specific time points.
- A perturbative scheme was developed for application to weakly correlated neurons.
Conclusions:
- The developed method offers an efficient way to perform second-order Volterra series reconstructions from spike trains.
- While overall second-order contributions are small, they are significant for specific stimulus types (rotational).
- The approach is scalable to larger neural populations, advancing the analysis of complex neural coding.
