Saccadic latencies are conventionally viewed as reflecting the accumulation of information during decision-making process (e.g. Carpenter and Williams, 1995) but we have previously shown that latency distributions may be strongly affected by reinforcement contingencies (Madelain et al., 2007). Here, we probe the possibility to control saccadic latencies in a choice paradigm.
Six subjects made saccades within 80-300 ms following a target stepping horizontally by 10 deg between two fixed locations. For each subject we selected two classes of latencies, ‘short’ and ‘long’, using the first and last quartiles of the baseline distribution (e.g. [80;151] ms and [185;300] ms respectively). We then concurrently reinforced each class in three blocked conditions (approximately 20000 saccades per subject) with different probabilities such that the relative frequencies of reinforcing ‘short’ versus ‘long’ latencies were either 9/1, 1/9 or 1/1.
Latency distributions varied depending on the experimental conditions: distributions shifted toward the shorter or longer values or became strongly bimodal. Moreover, the relative proportions of ‘short’ and ‘long’ latencies matched the relative proportions of reinforcers earned from each option (slope up to 0.95). We used the LATER model (Reddi and Carpenter, 2000) to investigate which parameters might best explain these changes in latency distributions.
Our results reveal that learned contingencies considerably affect the allocation of saccades in time, and provide strong evidence for a voluntary control of saccadic latency. The functional significance of this control extends well beyond information accumulation.