ISTOCK
HSE researchers have developed an algorithm for estimating individual response preparation period duration. Their approach can help diagnose disruptions in decision-making and motor functions associated with certain diseases. The study findings are published in PLoS ONE. The research was financed by a megagrant from the Russian government as part of the 'Science and Universities' National Project.
The speed and accuracy of one’s responses to an ever-changing environment are critical for adaptation. Generally, the more predictable an event for an individual, the faster and more accurate their response. A cue immediately preceding it can make an event more predictable. A yellow traffic light, for example, is a cue that you will soon be allowed to move forward. But you can only move once the light turns green, which is a stimulus for action.
Scientists outline three phases that take place between a stimulus to action and the action itself:
The decision-making phase can also occur after the cue and before the imperative stimulus, as in the yellow and green traffic lights example. This phase is then called the preparatory period.
The traffic light example also shows that a cue preceding an imperative stimulus reduces the reaction time by replacing decision-making with the preparatory phase. The effect of the preparatory period on motor reaction has been confirmed by experiments: the more time is spent on preparation, the faster one responds to the imperative stimulus.
Experimental tasks involving a cue make it possible to estimate the decision-making time (the preparatory period) separately from the motor reaction time. However, until recently, researchers did not know how to accurately measure the preparatory period, which was pre-set manually in such experiments. HSE researchers developed an effective algorithm for measuring the preparatory period duration in each individual case.
The experiment used a classic forced choice response task with a preceding fully informative cue about response direction. In each trial, the participants were presented with a certain figure as a cue and were instructed to click on either the right or the left arrow on a keyboard depending on the its shape, but only after the cue disappeared and was replaced by a green circle (the imperative signal to action). The time between the cue and the imperative signal was the preparatory period in the decision-making process.
Each subject performed 120 trials during the experiment. The preparatory period varied across trials:the cue appeared over the green circle in 40 trials, in another 40 trials it appeared after 1200 ms, and in the remaining 40 trials the period between the cue and the signal was adaptive, ie adjusted proportionally to the individual duration of subject's response period in previous trials.
The experiment involved a total of 67 subjects, who were divided into two groups and given either an easy or a hard task. The cues presented to the subjects in the easy version differed only in shape (either a square or a cross), while the hard version required the subjects to consider both the shape and the rotation angle. As expected, the latter group took more time to decide which arrow to choose.
After the experiment, the researchers examined the reaction times in both groups based on the task difficulty and trial conditions. Where no preparatory period duration was allowed, ie the signal and cue were presented simultaneously, the time between the signal and the subject's response was longer in the 'hard task' group compared to the 'easy task' group. While indicating that the hard task required more time for decision making, this finding did not allow for an estimation of the decision-making time separately from the motor reaction time.
Where the preparatory period was 1200 ms, the reaction time was the same in both groups, indicating that all decision-making was shifted towards the preparatory period. However, this still did not give the researchers a clear idea of individual decision-making time.
Our analysis of trials using the adaptive algorithm reveals that the preparatory period was longer in the 'hard-task' group, while the reaction time was the same in both groups. Thus, only by using the adaptive algorithm could we estimate the actual decision-making time, which in this case varied depending on the complexity of the task.
Unlike other approaches to measuring decision-making time, the adaptive algorithm estimates it during the task completion rather than after it. In addition, this algorithm makes it possible to manipulate the process of completing the task.
This approach can be used for scientific research of decision-making and also for practical purposes, eg to diagnose impairments in decision-making and motor functions associated with certain diseases. Some conditions, such as anxiety or schizophrenia, are known to affect decision-making but not motor function; some others, such as Parkinson's disease, tend to disrupt motor function but not decision-making, while still others, such as stroke, can damage both. The proposed algorithm makes it possible to assess each of these two aspects individually by measuring a patient's motor reaction time separately from the decision-making time.
Work is underway to pilot the new method in a clinical setting at the Laboratory of Medical Neural Interfaces and Artificial Intelligence, which was created at the Federal Brain and Neural Technology Centre jointly with the Pirogov Russian Research Medical University as part of the Priority 2030 programme.
IQ
Text author: Anastasia Lobanova
Alexey Tumialis
Study director and author of the methodology; Leading Research Fellow at the HSE Centre for Bioelectric Interfaces