Abstract
Recent work has indicated that adaptation is controlled by multiple mechanisms acting at different timescales, but their neural underpinnings remain largely unknown. Here we explored this issue with two psychophysical experiments. We first used a “deadaptation” procedure (Bao, Fast, Mesik, & Engel, 2013) to investigate contrast adaptation under either the binocular or monocular adaptation condition. Thresholds were measured with a spatial 4-AFC contrast detection task, tracked by a one-down-one-up staircase. The “deadaptation” duration was individually set, based upon a pilot test where 5 minutes of adaptation to high contrast was followed by 280s of deadaptation. Interestingly, the effects of adaptation failed to completely decay to the baselines under the monocular condition in 21 out of 37 subjects, but succeeded under the binocular condition in everyone. For the 16 subjects whose adaptation’s effects could deadapt to the baselines, their data showed spontaneous recovery in the post-tests, demonstrating that multiple mechanisms controlled adaptation in both conditions. Critically, it took longer to deadapt to the baseline (p < .001) in the monocular (84s) than the binocular condition (51s). Given the larger proportion of monocular neurons activating in the binocular condition, these results suggest that longer-term mechanisms reside at the binocular processing stage. To compare the timescales of mechanisms spanning wider in the visual processing hierarchy, we modified Hancock and Pierce’s (2008) paradigm to track the decay of TAE in 15 subjects after adaptation to either the compound gratings (curvature) or component gratings. An exponential function was fit to the timecourses following adaptation, which revealed slower time constant (p < .05) for the compound (149s) than the component (113s) adaptation condition. This indicates longer-term mechanisms in the mid-level (e.g. V4) than the early visual areas. Our findings imply that cortical mechanisms for controlling adaptation may become more sluggish along the visual processing stream.
Meeting abstract presented at VSS 2015