Category: Brain Science

Those who read my blog post a couple days ago (“Information that Moves from the Eyeball to the Brain“) will recall that I was not happy with the paper I reviewed. It seemed to me to represent a lot of work and presented no new ideas or findings. Then I noticed that another paper was published back-to-back in the same issue of the Journal of Neuroscience on the same topic. Would this paper enlighten me and show me the error of my earlier reading?

The paper titled “Recoding of Sensory Information across the Retinothalamic Synapse” published October 13, 2010 in the Journal of Neuroscience was by a different research team. While the experimental techniques were mostly the same as in the other paper, the analysis and discussion were quite different and, I thought, provided some useful new information. Interestingly, this paper clearly stood on the shoulders of a 1998 paper from the principle investigator that produced the paper I reviewed a couple of days ago.

This paper asks if pairs of spikes in retinal ganglion cells separated by certain time spans (between about 2.5 to 30 milliseconds) together carry information not carried by the individual spikes alone. If yes, is that information encoded and represented in single spikes in the receiving thalamic neurons. Their answer was yes to both questions.

Note: Those who need basic background on the visual system should consult my previous post “Information that Moves from the Eyeball to the Brain.”

They conclude that their “work provides a first biological example of the transformation of a correlation code into an independent code across a synapse through which sensory information flows from the periphery to the brain.” Their unique contribution was a joint encoding model that enabled them to show that thalamic neurons inherited selectivity for the spatial features encoded by their retinal inputs but gained sensitivity to different temporal features.

I found this paper satisfying because they looked at how information distributed in spikes in the retina was re-presented in the thalamus. They also provided evidence for synaptic integration underlying the transformation between retinal ganglion cells and thalamic neurons.

Other related blog posts:

Information that Moves from the Eyeball to the Brain

The recent paper “Decorrelated neuronal Firing in Cortical Microcircuits” published January 29, 2010 in Science takes a new look at correlated response fluctuations among simultaneously recorded neurons in the cerebral cortex.

This convincing study used very high quality simultaneous recordings from nearby neurons in the primary visual cortex to overcome technical challenges that may have contributed to a large body of literature that concludes that nearby neurons share significant common input based on correlated trial-to-trial variability in activity.

They saw correlations near zero using a number of different stimulus protocols and experimental conditions. They then went on to artificially introduce into their experiments some of the possible confounding factors and reproduced the significant correlations observed in earlier papers. The authors conclude that either 1) “adjacent neurons share only a few percent of their inputs” or 2) “their activity is actively decorrelated.”

At the back inside of each of your eyeballs is a sheet of nerve tissue known as the retina. Signals move from the retina to the rest of the brain through only one kind of nerve cell (neuron) known as the retinal ganglion cell.

Retinal ganglion cells communicate with the rest of the brain through a structure known as the thalamus and, more specifically, the part of the thalamus known as the lateral geniculate nucleus. So brain signals, known as spikes or action potentials, related to vision originate in the eyeball’s retina and are sent by retinal ganglion cells to the lateral geniculate nucleus of the thalamus from where they may travel to other brain structures.

A new paper asks if spikes in the retina that carry the most visual information are being preferentially funneled through the thalamus. The research team also looked at the role that spike timing may play in the information transfer process from eyeball to brain.

(The paper “Spike Timing and Information Transmission at Retinogeniculate Synapses” was published October 13, 2010 in the Journal of Neuroscience.)

The authors define a spike in a retinal ganglion cell followed after a short period (around 2.5 milliseconds) by a spike in an associated thalamic neuron as a relayed spike. Only a subset of spikes in the retina are relayed to the thalamus.

The conclusion reached in the paper was that spikes carrying the most information were selectively relayed from the eyeball to the brain. They base this conclusion on the following results:

• relayed spikes had a stronger correlation with the visual stimulus than non-relayed spikes.
• short intervals between spikes were more effective at driving thalamic neuron responses than spikes that followed longer intervals.
• relayed spikes occurred with significantly greater spike timing precision (less variance) than non-relayed spikes.
• relayed spikes had significantly less variance in spike number than non-relayed spikes.
• relayed spikes carried significantly greater information than non-relayed spikes.

While the techniques and analysis appear sound I find myself ill at ease with this paper and its conclusions. It seems to me that at least the first of these findings is circular. A receptive field is a foundational idea in the brain sciences. Sensory system brain cells exhibit receptive fields that by definition show greatest response and shorter latencies to stimuli in particular parts of sensory space. The second finding follows directly from fundamental aspects of our understanding of how neurons process input. And the last three findings are straight forward consequences of filtering out the spikes that are less relevant to the receptive field of the thalamic neuron.

Did we really learn anything new from this paper? Is the relay between the eyeball and the brain really acting like a filter or is there something more subtle going on where both “non-relayed” and “relayed” spikes are contributing to the code found in the thalamus? Do the authors present evidence to support the conjecture that retinal ganglion cells “appear to use discrete firing events as the symbols with which to encode the visual world?”

Am I missing something?