Information conveyance in the brain is thought to occur primarily in the form of action potentials. These electrical signals are the currency of the nervous system—for sensory reception and all subsequent processing—and are therefore commonly measured as to understand the activational responses of specific brain structures. While action potentials unquestionably represent an essential metric of cortical communication, the work of Dr. Priebe—of UT’s Neurobiology section—suggests that spike rates may provide an incomplete account of visual cortical behavior, and that subthreshold membrane potentials can provide an additional source of inference regarding the processing mechanisms of the mammalian brain. Dr. Priebe collects intracellular recordings from the primary visual cortex (V1) of cats, a brain structure that is innervated by neurons originating in the lateral geniculate nucleus (LGN) of the thalamus. A distinct characteristic of V1 cells, as compared to thalamic (or retinal ganglion) cells, is an increased degree of selectivity for stimulus features such as orientation, direction of motion, and depth. Two models have been proposed to explain this increase in feature-selectivity: (1) the feedforward model, and (2) the feedback model (lecture). Briefly, the feedforward model suggests that (solely) excitatory projections from LGN relay cells are sufficient for the observed specificity in V1 response patterns, while the feedback model contends that further refinement must take place in V1 in the form of lateral inhibition from differentially-tuned, nearby cells.While the accumulated evidence has mostly supported the feedback model (lecture), this more predominant explanation is unable to account for certain important findings, such as the mismatch between the shape of orientation tuning curves as measured by spike rates, and as predicted by the patterns of thalamic input arriving at V1; a linear model predicts a tuning function that is three times broader than the observed spike-rate data (Priebe and Ferster, 2008). This discrepancy can be resolved, the authors note, if spike thresholds, and other nonlinear properties of visual pathway cells, are taken into account. Specifically, their findings indicate that membrane potential responses are at times better predictors of the selectivity of V1 cells thanare spike-counts. When looked at from this perspective, several visual phenomena that had relied upon lateral inhibition within V1 for explanation can be accounted for sufficiently by appealing only to the feed-forward thalamic inputs into V1. The described work provides a reminder that action potentials are but one measure of brain activation, which may not provide complete agreement with receptive field properties or other measures of activation. Further work in this vein comes from experiments testing subjects with strabismus, a visual abnormality in which the two eyes are slightly offset. It was once widely believed that persons with this disorder did not integrate information across visual fields, that primary visual cortex (V1) was essentially monocular in its input connectivity (lecture). Rather than there being a stringent segregation of input from the two visual fields, however, data suggest that V1 cells do in fact receive information from both visual fields, but that the contralateral inputs are unable to elicit action potentials (lecture). What is clear, then, is that inferences about connectivity in the brain can be misconstrued if action potentials are the sole measurement tool. The work of Dr. Priebe is both interesting and important in that it engenders change in thepredominant way of thinking about specific visual phenomena. Even more generally, it encourages one to think about the role that subthreshold membrane potentials may have on a variety of suprathreshold responses. For example, complex cells in V1 rapidly transitionbetween so-called ‘up’ and ‘down’ states of activation in the absence of visual stimulation (lecture). V1 cells in different regions of cortex will show synchronous activity if they are tuned to similar stimulus features. What could be the purpose of this subthreshold ebb and flow? Said differently, what possible function might be accomplished by patterned neural activity that is insufficiently weak to trigger action potentials? We know that neurons travel great distances, are insulated as to reduce propagation times,show redundancy in coding, process information hierarchically, and are ordered in brain structures in near optimal ways—all this in an effort to best serve the organism. It also appears to be the case that the nervous system has reserved action potentials only for the strongest, most salient inputs. As Dr. Ballard noted, EPSPs are expensive in that they consume some 50% of a cell’s total energy (lecture). So perhaps it would be unsurprising that the brain might employ additional, more efficient methods of processing that do not require such a large proportion of its metabolic resources.Critics of Dr. Priebe’s work could cite that his measurements are intracellular, which are inherently more difficult to obtain than those made extracellularly. Furthermore, subthreshold potentials are likely to be less informative, in general, than those that are strong enough to produce action potentials. Even though certain information may be ignored with spike-rate measures, those who use action potentials as their primary metric may not care about this loss. Action potentials lead to the release of neurotransmitters, whose functions we largely understand.And we understand that electrical signals propagating from one structure to another are essential for the type of information encoding and processing that is required for sophisticated cognitive and perceptual functioning, and subsequent action. At the present time, it is much less clear how changes in subthreshold membrane potentials might facilitate communication between the structures involved in these roles. It is additionally difficult to say with certainty that changes in subthreshold potentials are predictive of meaningful functionality.In summary, cortical information is transmitted through action potentials, but evidence exists to suggest that more general levels of activation—as measured intracellularly—are important as well in understanding cortical processing. The power of Dr. Priebe’s approach is that it has the potential to
View Full Document
Unlocking...