Sequence-dependent sensitivity explains the accuracy of decisions when cues are separated with a gap

Abstract

Most decisions require information gathering from a stimulus presented with different gaps.Indeed, the brain process of this integration is rarely ambiguous. Recently, it has been claimed that humans can optimally integrate the information of two discrete pulses independent of the temporal gap between them. Interestingly, subjects' performance on such a task, with two discrete pulses, is superior to what a perfect accumulator can predict. Although numerous neuronal and descriptive models have been proposed to explain the mechanism of perceptual decision-making, none can explain human behavior on this two-pulse task. In order to investigate the mechanism of decision-making on the noted tasks, a set of modified driftdiffusion models based on different hypotheses were used. Model comparisons clarified that, in a sequence of information arriving at different times, the accumulated information of earlier evidence affects the process of information accumulation of later evidence. It was shown that the rate of information extraction depends on whether the pulse is the first or the second one.The proposed model can also explain the stronger effect of the second pulse as shown by Kiani et al. (2013).

A temporal hierarchical feedforward model explains both the time and the accuracy of object recognition

“…In other words, spikes generated in different time points which are indicators of the edge's strength and the strength of the pre-learned patterns in the input image, are not redundant and they can improve the categorization performance. This improvement in performance as a result of longer stimulus duration, is in line with results in behavioral studies 29 . It should be noted that, in this analysis, crossing the bound cannot be the decision rule because we forced the model to decide about input stimulus at different time points, no matter if any of the accumulators cross the bound or not.…”

“…The input category will be recognized as soon as an accumulator reaches a threshold. This accumulation to the bound mechanism of decision making is a well-known biologically plausible decision-making model 2,[27][28][29][30] . Results show that the temporal representation of the input stimuli is informative, since it can explain human reaction time as well as the accuracy, in an object recognition task.…”

Improvement of association between confidence and accuracy after integration of separate evidence over time

“…The two pulses separated by up to 1 s supported a level of confidence that was indistinguishable from a pair of pulses separated by no gap. Direct comparison between single-pulse and double-pulse trials, along with previous studies Tohidi-Moghaddam et al, 2019), showed that participants' accuracy significantly differed (t(11490) = -3.09, p < .05, 95% CI = [-.08, -.02], Cohen's d = .11). However, in double-pulse trials participants were not more confident comparing to single-pulse trials (t(11490) = 1.35, p = .18, 95% CI = [-.01, .06], Cohen's d = -.05).…”

“…Finally, how implicit markers of confidence -response-time, CPP and pupillometry-change after receiving separated pieces of information. Here, to clarify confidence, we required observers to make a two-alternative decision after viewing either one (single-pulse) or two (double-pulse) motion pulses separated by four various temporal gaps (similar to (Kiani et al, 2013;Tohidi-Moghaddam et al, 2019)). We performed several logistic regression models to measure the impact of stimulus characteristics on confidence.…”

“…In this scenario, as the number of pieces of information increased, the interpretation of the car's direction would be improved. Indeed, research has shown that the accuracy of decisions can be significantly improved by integrating This is a provisional file, not the final typeset article information from separate cues (Kiani, Churchland, & Shadlen, 2013;Kira, Yang, & Shadlen, 2015;tickle, Tsetsos, Speekenbrink, & Summerfield, 2020;Tohidi-Moghaddam, Zabbah, Olianezhad, & Ebrahimpour, 2019;Waskom & Kiani, 2018). Typically, our decisions are accompanied by feelings that reflect the likelihood that the decision is correct; such a feeling is called confidence (Kiani, Corthell, & Shadlen, 2014).…”