Antisaccades as decisions: LATER model predicts latency distributions and error responses

Abstract

Antisaccades are widely used in the study of voluntary behavioural control: a subject told to look in the opposite direction to a stimulus must suppress the automatic response of looking towards it, leading to delays and errors that are commonly believed to be generated by competing decision processes. However, currently we lack a precise model of the details of antisaccade behaviour, or indeed detailed quantitative data in the form of full reaction time distributions by which any such model could be evaluated. We measured subjects' antisaccade latency distributions and error rates, and found that we could account precisely for both distributions and errors with a model having three competing LATER processes racing to threshold. In an even more stringent test, we manipulated subjects' expectation of the stimulus, leading to large changes in behaviour that were nevertheless still accurately predicted. The antisaccade task is widely used in the laboratory and clinic because of the relative complexity and vulnerability of the underlying decision mechanisms: our model, grounded in detailed quantitative data, is a robust way of conceptualizing these processes.

A computational analysis of flanker interference in depression

“…Our computational model uses a modified version of the Linear Approach to Threshold with Ergodic Rate (LATER) (Noorani & Carpenter, 2013); the key modification is that noise in the accumulation process is explicitly modeled. This should not yield qualitatively different interpretations as these models are closely related (Bogacz et al, 2006; Donkin et al, 2011).…”

Using saccades to diagnose covert hepatic encephalopathy

“…1). It is far larger than can be explained by such factors as nerve conduction velocity or synaptic delay: saccadic reaction time essentially represents the time needed to make a decision as to whether a potential target is actually worth looking at (Carpenter 2000), and in recent years the increasing interest in making precise measurements of latency has led to important insights into the neural mechanisms underlying decision (Carpenter 1981;Carpenter and Williams 1995), (Noorani and Carpenter 2013;Schall 2005).…”

Basal ganglia: racing to say no

“…Once a particular decision-making unit reaches its threshold, it determines the outcome of the decision by triggering a response. Although initially used to explain simple decision-making tasks, such as looking towards a novel visual stimulus, it has recently been successful in explaining more complex decisions [3][4][5] especially in terms of predicting full reaction time distributions, a very stringent test of any neural model [6].…”

“…In this task, a subject is told to look away from, rather than towards, a suddenly-presented visual target, generating anti-saccades instead of the usual pro-saccades, and requiring suppression of the subject's natural tendency to fixate on the target. A recent study has shown [5] that human antisaccade behaviour can be modelled successfully as a race between three distinct decision units: Prosaccade, Antisaccade, and Stop ( Figure 1).…”

“…If the Stop unit wins, a prosaccade is cancelled and this allows the antisaccade to continue to threshold. In this way, the stop unit is important for cancelling many of the errors that would otherwise have occurred and this was shown to be essential for quantitative prediction of error and antisaccade response proportions [5]. By looking in detail at the resultant distributions of response times for the different outcomes (Figure 2), it is possible to comprehensively characterise the parameters of the models.…”