Notes from Breakout Session I

On the second day of the workshop the participants broke out into four different groups based on their research oreintation. The groups were (1) Pyschology and Psychophysics, (2) Anatomists, and (3) Theorists. Due to the large number of theorists, the theory group was split into two (Theory 1 and Theory 2).

The groups met separately for 45 minutes. Each group was supposed to come up with a list of three things that they want from the other groups. At the end of 45 minutes all the groups met at one place and a discussion ensued on the requirements. What follows is a compilation of suggestions from different groups.

Suggestions From Psychology/Psychophysics group

For theorists

• models that can act on real stimuli real images, moving images
• Are there associative memory models running on the hardware that we have? (spiking models)
• Theory of mid-level vision that is understandable to us.

For anatomists

• more quantitative descriptions organized into ciruits.
  proportions of neurons in one area compared to another area are they myelinated or not
• More oranized quantitative information.
• Timing
• Thalamus
• Anatomists comments on theory slides.

From Theory Group I

For Anatomists

• Naturalize Stimulii
• Make the raw data available.
• Fast-forward development of multi-electrode recordings.
• Specify data on excitatory/inhibitory..which layer they are in.

For Psychologists

• Is the BOLD signal reactive or predictive? Need explanation.
• Request for more mathematical/theoretical training.
• Theoretically driven experiments.
• Theories posited before data.

From Anatomists

For Theorists

• Falsifiable predictions
  crucial tests.
  - test on how bayesian inference occurs in brain.

• Rigorous comparisons between models.
  - surface representation (biologically plausible)
  - Where does the model fail?
• Different feedback systems.
  - predictions from these models.

To Psychologists

• Ken: What approach, what end point?
• More natural stimulii.

From Theory group II

For Both Anatomists and Psychologists

• Better communication
  - theoretical training (academic curriculum)
  - tutorials or invited talks @ conferences.
• Better Data
  A) centralized database, systematic connectivity diagram
  B) Reporting (availability/completeness)
  C) Natural stimulii/ multi-electrode recordings.
• Hire us!!

Back to the main index for Inference and Prediction in Neocortical Circuits.