Neuron Model Analysis in
Parallel coordinate display using MATLAB

Not even nearly ready yet!

Introduction

Inselberg () and Wegman () showed that a visualization system based on parallel coordinates could be useful in understanding high-dimensional data sets.

We are currently working with a 6-parameter system.

The Code

First try at parallel cood in Matlab.

Results

Each of the following images is based on 10,000 random parameter selections in 6-space. The parameter axes are:

  1. initial threshold
  2. membrane time const
  3. sensitivity to K conductance
  4. refractory time const
  5. threshold sensitivity
  6. threshold time const
All axis ranges are normalized to unity.
Cells falling into the "short" ISI bin (dF=8):


Cells falling into the "medium" ISI bin (dF=8):


Cells falling into the "long" ISI bin (dF=8):