Professor Emeritus Minoru Kimura, Professor Yutaka Sakai, and their colleagues of the Brain Science Institute at Tamagawa University, in a collaborative study with Institute of Science Tokyo, Fukushima Medical University, and Kyoto University, made use of genetically modified animals, optogenetics, neuroscience, and computational models to discover that the indirect pathway, one of the main circuits of the basal ganglia, plays a new function different from its previously known learning functions.
Kimura stated, "Humans and animals repeat actions that yield good results. However, sometimes they experience failures or strained relationships with others. To resolve this, it is necessary to actively learn from failures. The brain mechanism for this has been unclear for many years, but we have elucidated part of it." Their findings were published in Science Advances.
It is known that the direct pathway of the basal ganglia contributes to reinforcing desired behaviors, while the indirect pathway helps avoid undesirable behaviors. Through experience of actions and their outcomes, the direct and indirect pathways work together to acquire behaviors that maximize the results (rewards) obtained. In stable environments, optimized behaviors lead to desired outcomes, but when the environment changes, habitual behaviors can hinder new trial-and-error exploration.
This study discovered that the indirect pathway not only helps avoid undesirable behaviors but also plays a role in continuing to search for alternative strategies when actions that should be desirable fail to produce results.
In the study, rats were tasked with learning behaviors that yield desired results through trial and error in uncertain situations. With their heads fixed in place, rats could obtain water by either pushing or pulling a handle with their forelimbs (with 80% and 20% probabilities). The high-probability option switched after several dozen trials. Through trial and error, they had to change their choice between pushing and pulling. After training, rats were able to select the option with higher reward probability more than 80% of the time after 20-30 trials following the switch.
When examining neural activity in the indirect pathway during these behavioral tasks, the researchers found that directly after unrewarded suboptimal actions, neural activity in the indirect pathway increased for 200-500 milliseconds. Furthermore, when the lower reward probability option was chosen, sustained neural activity persisted for approximately two seconds afterwards. When the brief neural activity immediately following the no-reward cue was large, they often repeated different behaviors. This result is consistent with previously known activity characteristics of the indirect pathway. Conversely, when the sustained neural activity after a brief pause was large, repetition of the action that yielded no reward became more common. The relationship between this latter neural activity and behavior represents a control mechanism opposite to what was previously understood.
Using optogenetic methods, stimulating the indirect pathway during periods of sustained neural activity increased the persistence of low-value behavior seeking compared to when not stimulated; conversely, inhibiting these pathways reduced such exploration. Neural activity manipulation was found to affect choices up to 2-3 trials later.
The study demonstrated that the indirect pathway of the basal ganglia is involved in the function of trying alternative actions when desired actions fail to produce results, and in continuing to explore alternatives even when those also fail.
Sakai suggested, "This may be a function that accelerates the reward maximization system reflecting immediate results by incorporating not only the most recent outcome but also longer-term evaluations."
Journal Information
Publication: Science Advances
Title: Dorsomedial striatum monitors unreliability of current action policy and probes alternative one via the indirect pathway
DOI: 10.1126/sciadv.adt4652
This article has been translated by JST with permission from The Science News Ltd. (https://sci-news.co.jp/). Unauthorized reproduction of the article and photographs is prohibited.