Abstract
Purpose:
Near absolute visual threshold in humans and mice, a dim flash produces a photon absorption in roughly one out of every five hundred rod photoreceptors. Since rod bipolar cells contact about 20-100 rods, nearly all responses depend on the reliable synaptic communication between one rod and one bipolar cell. Our lab has demonstrated multiple links between genetic manipulations and synaptic physiology in the regime of single photon detection, and in one case, an effect on behavioral performance. Our goal was to design a protocol for measuring absolute visual threshold using a high-throughput, low-maintenance, non-invasive, cost-effective method that mice could learn quickly, and which was robust to differences in task aptitude.
Methods:
Wild type mice were trained to respond to a brief LED flash with a head poke, which interrupted an infrared beam between their small enclosure and a water source. The percentage of correct detections was measured for a range of flash strengths, yielding 300-400 trials per day on an average 45 minute experiment. To estimate how often the mouse was attending to the task, we determined the flash strength at which percent correct did not improve with increased brightness. On some experiments, we used dummy flashes to determine the percentage of trials that were scored correct as a result of a lucky guess. These percentages were used to estimate the probability of a correct trial when the mouse was attending to the task, and the flash strength that produced 50% correct was taken to be threshold.
Results:
The mice learned the simple task described in less than a week. Additional acclimation was required for performing the task in darkness, and following prior dark adaptation. The mice attended to a majority of trials (65-72%, N=2), and guessed correctly on a minority of dummy trials (10-25%, N=2). Guess probability was greater when a dimmer range of flashes was presented. Thresholds for both mice tested were not significantly different.
Conclusions:
Preliminary results suggest that our method produces a robust estimate for absolute behavioral threshold, and that mice have the ability to learn the task in a reasonable amount of time. More effort is needed to adjust experimental parameters for optimal learning speed and task performance, but the protocol in its present form would be sufficient to detect small differences in threshold between different populations of mice, and over their life span.
Keywords: 688 retina •
649 photoreceptors: visual performance •
641 perception