Why We Are Responding Slowly, When There Is A Decision Conflict?

One may usually think that why we are responding slowly when there is some decision conflict. One may think that it is a mechanism for getting more evidence since we have a conflicting decision and this suggestion is completely logical. But, what could be the possible mechanism behind that? Anything that is happening in our environment, or in our reality will have a representation in our brain. Then how the decision conflict is represented in the brain? How it is regulating the decision-making process? What parameter of the noisy accumulation of evidence is being changed by this process? Cavanagh et al (2011) seem to be a perfect place to search for these questions. There is a PFC-basal ganglia system that controls the decision-making process when subjected to a difficult decision-making task. That is it is generally observed that in general, in such situations participants would need more time to respond. What would be the possible mechanism behind this? It is like only when control is needed, mPFC is interacting with STN. There are projections from mPFC to STN. This system seems to be involved in regulating habitual responses. This system particularly has regulatory action when cortico-striatal signaling results in impulsive response. Possibly, there could be signals that tell how difficult a particular task is, and it is possible that this system is additionally activated only when the conflict is high and not when it is low. Even though there was evidence for such a hyper direct pathway, much of it had remained correlational. Then, what is the possible way to test this hypothesis? One way is to disrupt the function of STN by high-frequency deep brain stimulation(DBS). But, what could be our test and control groups? It turns out that STN DBS is a treatment method for Parkinson’s disease. That is if we induce DBS, it is possible that the decision would be more impulsive. It disrupts the slowing down when faced with a difficult decision. The proposed hypothesis is that the mPFC-STN system modulates the process such that, the decision threshold would be high for difficult tasks, but when we use DBS, this process is interrupted.EEG could be used to assess the activity of mPFC during conflict trials. Theta power increases following error, punishment, or conflict. And this quantity seems to have a linear relation with the slowing of reaction time. They manipulated or directly measured the activity of STN. Both healthy and individuals affected by Parkinson’s disease performed reinforcement learning and conflict decision task and the EEG was recorded. Here, it was interesting to see that both cortical theta, response time, and the threshold were rising during response conflict. But interestingly the relation between mPFC theta ad decision threshold just inverses on applying STN-DBS. Thus it is suggestive that STN modulates the cortico-basal ganglia responses when subjected to a conflict. Firstly, they looked at the learning of stimulus reinforcement probability, in ON, and OFF DBS trials, when the participants are involved in high conflict or low conflict trials. But from a t-test, it was confirmed that there is no significant difference between the two conditions. But, one can actually see the high conflict trials as a percentage change from low conflict trials, and this, in turn, could be a function of accuracy. That is lower conflict and high RT could suggest that there is a high probability of that response being correct and vice versa. When ANOVA is repeated for valence(HC, LC), DBS condition, and accuracy(optimal, suboptimal), it is observed that there is a significant difference between valences and accuracy and for the interaction between these two. RT was faster for Win-Win trials and for suboptimal accuracy and for conjecture between these two. What does this say? Win-Win trials are HC trials and RT is faster and accuracy is low, which in turn suggests that the presence of impulsive response when subjected to conflict, which is what expected. But, for those who had Parkinson’s disease, it is observed that they responded faster for both kinds of HC trials, but with suboptimal accuracy, when DBS was on. But, when we see the OFF DBS trials it is seen that the valance has more influence than accuracy. Ths from this it is observed that HC suboptimal responses are seen in deceased individuals when they are in the ON DBS trial and not in OFF DBS trials. 

They actually wanted to determine whether the trial to trial mPFC theta is predictive of the RT delay or whether there is any influence of STN-DBS on that. Computed theta-RT values for HC and LC trials. The weights for this relationship were estimated using linear regression and larger weights are indicating a strong correlation. 

   (a) Stimulus presentation and response commission were characterized by notable beta power suppression and theta power enhancement compared with baseline in both ON and OFF conditions, which were combined here. (b) Topoplots of the high-low conflict difference in standardized regression (β) weights for cue-locked theta power and response time (±0.1 std β). The FCz site is indicated on the control topoplot. (c) Standardized regression (β) weights (mean ±s.e.m.) for cue-locked theta power and response time, demonstrating that DBS reversed a natural coupling of theta band power with response time slowing on high-conflict trials.

Adapted from: Cavanagh et alCue-related mPFC theta power predicts slower RT for both healthy and DBS OFF diseased individuals, but, not in DBS ON affected individuals.   There was a significant difference between DBS conditions during high conflict trials. There was no difference between the two high conflict trials also. But additionally one has to confirm that the EEG pattern observed is indeed a feature of the human brain and not due to the deceased condition. For that, the diseased individuals in ON and OFF trials are compared with the healthy individuals. And it is observed that the healthy ones which is the control sample don’t differ much from the diseased ones in DBS OFF. But, in DBS ON trials the behavior was significantly different. Simply mPFC theta shows a positive correlation with response time and hence slower response when the response is high conflicting. On the other hand, when we disrupt the mechanism with STN DBS, it is observed that this relationship just reverses. 

Drift Diffusion Modeling

DBS ON/OFF study: Bayesian posterior densities of decision thresholds estimated from the drift diffusion model (ordinates) and how they varied as a function of mPFC theta (abcissa). Peaks of the distributions reflect the most likely value of the parameter. Significance was assessed by at least 95% of the distribution being to the left or right of zero. (a) Simple effects of theta. OFF DBS, increased theta was associated with increased decision threshold for high-conflict trials, but not low-conflict trials. ON DBS, increased theta was associated with a decreased decision threshold on high-conflict trials, but not low-conflict trials. (b) Theta × conflict interaction. Increases in theta resulting from high > low conflict were associated with increases in threshold OFF DBS and in healthy controls. The opposite pattern was seen ON DBS. (c) These threshold effects are reflected by changes in response time distributions. These plots show the best fit response time distributions for optimal and suboptimal choices as a function of low and high mPFC theta/threshold in affected individuals in OFF DBS sessions. Higher theta power is associated with a reduction in the density of fast suboptimal choices and greater dispersion of optimal response time distributions, fitting with an account of increased threshold.

(adsbygoogle = window.adsbygoogle || []).push({});

Adapted from: Cavanagh et alUltimately what they wanted to see is, whether the hypothesized effect of mPFC-STN on decision threshold is significant, or to see whether we can get such a relationship. Notice that in DDM, we can only handle right and wrong responses, which are optimal and suboptimal trials here. When the stimulus is presented, the response from mPFC would be mainly dependent on the prior probability. But when there is response conflict, the STN-mPFC increases the decision threshold such that the agent would get more time to get to the final decision and the amount of evidence accumulation would be also higher. This could be for getting more time for evaluating the reward values. HDDM, in this case, uses the hierarchical Bayesian method to get the posterior distribution for all the model parameters. This allows us o simultaneously calculate the parameters for the individuals and the groups, allowing it to vary between individuals in a group. Here the regression coefficient for theta power in the mPFC and decision threshold for LC and HC trials are obtained this way. And they also looked at whether it was affected by the interaction with DBS status. It is noticed that for deceased individuals the mPFC theta and decision threshold has a strong positive correlation in high conflict trials, and that effect wasn’t present in low conflict trials. As we had already said the relationship was just reversed when DBS is turned on. This was true for both healthy and diseased classes for which the condition was DBS OFF. A higher threshold means we need more time for accumulating evidence and there would be fewer suboptimal decisions taken in the process as more evidence is being accumulated. 

(adsbygoogle = window.adsbygoogle || []).push({});

Intracranial EEG from the STN for dorsal, middle and ventral leads. Both beta suppression and theta enhancement were observed in the STN. The rightmost columns show the condition-wide differences revealed by permutation testing. High-conflict trials were characterized by a diminishment of low-frequency power across leads, greater post-cue activity in the dorsal lead and greater post-response activity across leads. The bottom row shows intracranial EEG data filtered from 2.5–4.5 Hz

Adapted from: Cavanagh et alThey also found that decision conflict is actually reflected in local STN activity in the same period that is observed in the mPFC activity, and a similar effect was observed for both the HC trials. 

Finally, what is shown in the paper is that the mPFC-STN system effectively represents the conflict between decisions and the DBS condition disrupts the ability of STN to react when there is a decision conflict, which is an effective rise in the decision threshold. Thus the decisions are consistent with the hypothesized model and hope the reader got the answer for the questions paused at the beginning. 

References:

[1] Subthalamic nucleus stimulation reverses mediofrontal influence over decision threshold, James F Cavanagh , Thomas V Wiecki , Michael X Cohen, Christina M Figueroa, Johan Samanta,

Scott J Sherman & Michael J Frank

(adsbygoogle = window.adsbygoogle || []).push({});

[2] HDDM: Hierarchical Bayesian estimation of the Drift-Diffusion Model in Python, Thomas V. Wiecki, Imri Sofer and Michael J. Frank