Energy Efficient Neural Codes

Abstract

In 1969 Barlow introduced the phrase "economy of impulses" to express the tendency for successive neural systems to use lower and lower levels of cell firings to produce equivalent encodings. From this viewpoint, the ultimate economy of impulses is a neural code of minimal redundancy. The hypothesis motivating our research is that energy expenditures, e.g., the metabolic cost of recovering from an action potential relative to the cost of inactivity, should also be factored into the economy of impulses. In fact, coding schemes with the largest representational capacity are not, in general, optimal when energy expenditures are taken into account. We show that for both binary and analog neurons, increased energy expenditure per neuron implies a decrease in average firing rate if energy efficient information transmission is to be maintained.

Highly Energized Inhibitory Interneurons are a Central Element for Information Processing in Cortical Networks

“…50,51 56,86,98,112 This finding underlines the general principle of sparse coding by pyramidal cells, which has been suggested to provide an optimal trade-off between information processing and energy efficacy. 113,114 However, pyramidal cells comprise B85% of all neurons in the hippocampus proper, and even with sparse coding they are able to encode multiple representations of spatial and episodic memories. 18,20,23,85 By propagating specific activity patterns into downstream cortical and subcortical regions, pyramidal cells also impact on network activity and information processing in multiple other brain regions.…”

“…It is likely, however, that patterns of neuronal network activity represent a compromise between maximal information processing and minimal energy utilization. 113,114 It is therefore important to understand the activity-dependent neuroenergetic demands and constraints of this highly active class of neurons. Indeed, several findings indicate that fast-spiking interneurons use much more energy than other cells in the central nervous system, which might render them particularly vulnerable to conditions of energy deficiency.…”

Theoretical understanding of the early visual processes by data compression and data selection

“…The output channel capacity H(O) = I(O; S) + H(O|S) wastes a fraction H(O|S) = H(N) on transmitting noise N which is typically less redundant between input channels, costing metabolic energy to fire action potentials (Levy and Baxter 1996). To minimize this waste, a different transform K is desirable to average out input noise, thereby introducing some redundancy in response O, i.e., correlation between different response channels and unequal use of different response states within a channel.…”

Vision and the statistics of the visual environment

“…Several authors have tried to incorporate metabolic costs as a constraint [54,55]. Recently, Balasubramanian and Berry [56] demonstrated that retinal ganglion cells in tiger salamander are optimized to transmit visual information at minimal metabolic cost, assuming the symbols of the neural code are represented by spike bursts of a given length.…”