No: Murray has been thinking about thalamus the wrong way. He writes:

“Both the thalamus and the olfactory bulb are the final stage in sensory processing before reaching target cortical regions”.

@PessoaBrain @WiringTheBrain @KordingLab @JaumeTeixi

1/3

First, his version of processing is simple neuromodulation and state selection; at most a bit of room for attention. No matter how you cut it the view of lemniscal thalamus as a dumb relay lives on. In his view there is frankly no change in representation or real processing.

2/

His dominance in this field with Guillery has held back alternatives of thalamic “relay” function sketched out by Adam Sillito in the 1990s.

The key to lemniscal thalamus is getting timing right—not modulating representation. Timing in optic tract is a disaster of noise.

3/

The way to parasynchronize noisy lemniscal input is to send a fast (magno) “abstract” of timing noise to layer 6 of cortex and then receive temporal feedback corrections from layer 6 that embed many learned priors. Finally LGN sends the temporally adjusted signals to layer 4.

4/

In this view, lemniscal thalamus is essentially a distant layer 7 of cortex that with layer 5 and/or 6 performs a global timebase correction for the benefit of neocortex AND non-lemniscal thalamus.

It surprises me how little neurophysiologist think about how brain achieves

5/

integrated temporal precision for accurate behavior. This is just as hard of a compute as is representational processing. Without parasynchronized timebases across systems signals cannot be chunked in virtual time for accurate compute both within and between subsystems.

6/

Time and space are relative in CNS and this is obviously true in the case of high spatial frequency vision. To dynamically create the appearance of visual world spatial stability at high resolution while moving takes a lot of temporal autocorrection.

7/7