Ecologists and physiologists assume that biological systems show a balance between energy costs and performance benefits, with natural selection acting to maximise performance and minimise costs. To see how performance varies with energy cost, Jeremy Niven from the University of Cambridge and two co-workers studied the activity of photoreceptor cells in four species of flies: the large, big-eyed blowfly Calliphora, the fleshfly Sarcophaga,and two much smaller, small-eyed species of Drosophila, including the smallest fly tested, D. melanogaster. Because photoreceptors in the four fly species are very similar, and their biochemistry and physiology are well understood, the team could reliably compare their performance.

By stimulating fly photoreceptors with a range of light intensities and by recording photoreceptor spike rate, Niven and his colleagues calculated that each species processed information at the same rate in dim light levels, 1000 photons per second, but that differences soon appeared as light intensity increased.

At the kind of intensity seen in broad daylight, 106 photons per second, the four species showed highly significant differences in processing rate, which were positively correlated with the size of the fly and of its eyes. Sarcophaga showed the highest processing rates, while D. melanogaster showed the lowest. Because Drosophila species fly slowly and tend to be most active at dawn and dusk, while Calliphoraand Sarcophaga are active in full daylight and fly fast, the authors suggest that these differences enable the larger, faster-flying species to process more information, more rapidly.

Using a simple electrical model of a photoreceptor, the team then indirectly calculated the rate of ATP consumption. They found that the energetic cost of maintaining these different coding capabilities was greatest in the species with the largest eyes and lowest in D. melanogaster,suggesting that as processing power increases in more intense light, so too does the energetic cost of producing that response.

However, the four species also differed in how much ATP they needed to keep their photoreceptors functioning at rest, in the dark. This substantial fixed cost contributes to the total processing cost, so to determine exactly how much it cost each species to signal a response, the authors subtracted the resting costs from the total energetic costs. The authors found that the cost of processing each bit declined with increasing light intensity for each species. But when they looked at the maximum processing power and compared it to cost per bit, they found that costs per bit were higher in the large-eyed Sarcophaga and Calliphora, which could process the highest amount of information, than in Drosophila, which could process the least. As a result, the higher the maximum rate of information that could be processed, the greater the costs of processing and signalling that information.

Although the large flies could process five times as much visual information as the smaller species, this increase in performance was coupled with a 25-fold increase in energy consumption, suggesting that there is a law of diminishing returns affecting processing power in the insect retina.

This elegant study not only confirms a key assumption of neurophysiology,that performance is balanced by cost, but also shows how relatively simple measures can be used to investigate complex aspects of the evolutionary and physiological forces that shape nervous systems. This approach might be usefully applied to studying processing power in other sensory systems, such as olfaction, and as a tool for investigating how an organism's environment can influence its nervous system's structure.

Niven, J. E., Anderson, J. C. and Laughlin, S. B.(
). Fly photoreceptors demonstrate energy-information trade-offs in neural coding.
PLoS Biol.