Article activity feed

  1. Evaluation Summary:

    This paper is of potential interest to psychologists and cognitive neuroscientists studying learning, decision-making, belief formation and metacognition. The authors use a clever, elegant task in which people make decisions with or without control over the information they sample, and link the cognitive processes at play to MEG and pupillometry signatures. The key finding is that when participants have control over information sampling (i.e. are seeking information), they need more contradictory evidence in order to switch their choices, and such switches are made with lower confidence. Anticipatory suppression of alpha-band activity in occipital and frontal regions occurred prior to decision switches, while pupil dilation increased post-switch. The authors propose a computational model to account for behavioral differences between conditions. However, some of the conclusions may not be fully supported by the data and alternative interpretations are possible; therefore further analyses are required to bolster the authors' claims.

    (This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. Reviewer #2 agreed to share their name with the authors.)

    Was this evaluation helpful?
  2. Reviewer #1 (Public Review):

    Roualt et al. use a task where people make decisions (and confidence judgments) with or without control over the information they sample. They show that when controlling the information sampling, people's choices are more 'sticky' due to a lower perceived hazard rate of the stimulus sequence. To account not just for choices but also confidence judgments, they extend a Bayesian model of hidden-state inference with a metacognitive module. They also link switches (and to some extend, model-derived changes-of-mind) to specific MEG and psychophysiological signatures.

    The conclusions of this paper are mostly well supported by data, but some aspects of the effects reported, the model and psychological interpretation need to be clarified and extended.

    Strengths:
    - the task is very well designed, with maximal similarity on perceptual and motor demands.
    - the data are beautifully presented, and the effects are very clear.
    - the topic of active sampling and exploration is of high interest, and it is timely to explore this with new experimental paradigms that improve on some aspects of previously used tasks.
    - using confidence judgments is a powerful way to disambiguate and extend computational models of behavior.

    Weaknesses:
    - quite a few different effects are presented, and it's not always clear which ones are novel (going beyond replications of Weiss et al. 2021); which ones are the crucial effects to disambiguate 'information seeking' and 'changes of mind' from potentially confounding factors; and which effects are are, and are not, captured by the Bayesian model.
    - the control experiment on pro- vs. retro-active decision-making is crucial, but without comparing these data directly (and statistically) to the main experiments it's hard to know whether this potential confound really plays no role in behavior.

    Was this evaluation helpful?
  3. Reviewer #2 (Public Review):

    Ps were presented with oriented bars from a blue or orange category, corresponding to CW or CCW tilt from the vertical but with variation around these orthogonal means. They are presented with 2-8 bars on each trial and asked to give the category from which they're drawn (cue-based trials, Cb), or to generate bars from one of the two categories (outcome-based trials, Ob). The category (blue vs orange) remained the same for several trials in a row and changed unpredictably. The change in Cb was simply that the presented category changed, but in Ob that the action-category mapping changed. Ps are slower to detect when the orientation changes in Ob trials and exhibit lower confidence in their decision, purportedly demonstrating that controllability increases the stability of beliefs. Changing your mind is also associated with stronger alpha suppression in the dorsal attention network, and followed by increased pupil dilation when presenting outcomes.

    The manuscript is well-written and interesting, with beautiful figures. I also found interesting that the alpha effects preceded the dilation effects. If supported, I would find the core conclusion (controllability increases the stability of beliefs) especially fascinating. However I was sceptical, and think the authors need to justify this conclusion on the basis of my concerns. I would like to believe it but can't see past an apparent confound. I hope the authors can reassure.

    Specifically, the Ob trials require selection of an action, monitoring the action-outcome relationship, and a judgement about the stimulus. The Cb trials, in contrast, I think simply require monitoring the stimulus (I could not figure out whether they perform an action to start each trial, but regardless, they will not need to keep track of action-outcome mappings). Therefore the fact that Ps are slower to notice the shift in Ob trials and that their judgements are associated with lower confidence would likely appear driven by the fact the Ps have three tasks (or at least two) rather than one. Relatedly, can the fact that alpha suppression and pupil dilation effects are increased with changes-of-mind in the Ob condition also be explained by the fact Ps have more tasks? If there are greater executive / working memory demands, this will reduce alpha and increase dilation (as already discussed in the manuscript). You detect stimulus changes less readily (etc) because there is a greater executive/working memory load in this condition, and this has nothing to do with controllability/action or stability of beliefs per se.

    I was somewhat reassured to see that Ps were claimed to be "equally sensitive" to the available objective evidence across conditions, but I equally could not see how this conclusion could be reached. It would be good for the authors to clarify what they mean here - did they test for a significant slope/precision effect in the data shown in Fig 3A? Is it possible to say Ps are equally sensitive to the evidence other than on switch trials, when the insight into their sensitivity is given by a response that only requires them to process the evidence on switch trials? But more generally, I think the reader needs convincing that the Ob-Cb differences are generated by something other than number of tasks and generic WM/executive differences, however the authors can do it. E.g., would the effects be equivalent if giving people an additional task to the visual that isn't motor? E.g., implicitly counting backwards; monitoring an auditory stream.

    My only other main point is that there is no discussion of alternative ways in which these data may have turned out. The discussion entirely reports claims and findings with which the present data are consistent. Are there no data or theories that could have led to alternative predictions? Currently the manuscript could give the flavour of such high consensus that there was little point running the studies - the empirical patterns to date all point in identical directions. I know this is not the case. One possibility that came to mind is that the stimulus ISIs are not very variable. 2Hz+-500ms. Given you can perceive better at certain oscillatory phases of visual processing (peaks), could Ps choose to start trials in Ob to align with peaks and thereby improve perceptual acuity? If this occurred, the influence must be weaker than those pulling in the opposite direction, but it could have led to an alternative outcome. If it's a quick analysis to perform/report and the authors agree with me, it may be interesting to see how the oscillations align with events differentially in Ob and Cb.

    Was this evaluation helpful?
  4. Reviewer #3 (Public Review):

    In this study, Rouault et al measure changes-of-mind and confidence variations in a perceptual decision-making task where information sampling is done with or without control. When subjects have control over information sampling (i.e. are seeking information), the authors found that they need more contradictory evidence in order to switch their choices, that such switches are made with reduced confidence and yet are more likely to be confirmed later on (rather than reverting back to the original choices). MEG and pupillometry data additionally revealed that anticipatory suppression of alpha-band activity in occipital and frontal regions occurred prior to decision switches, while pupil dilation increased post-switch, and these two neurophysiological responses were stronger in the controllable condition.

    This study comprises 4 experiments, one of which was previously published (Weiss et al, 2021, Nature Communications). The presence of multiple experiments and replicated findings help mitigate the concern of low sample sizes in each of the experiment. The main novelty of the study is that it includes measures of confidence and a sophisticated Bayesian model making robust predictions about both decisions and confidence ratings.

    While this work is likely to be useful and significant for the field of decision-making research, there are also several weaknesses and concerns that the authors will hopefully be able to address.

    1. One main concern is about the interpretation of the results - in particular, whether the observed differences between the cue-based and outcome-based conditions could be better explained by the presence of a target-induced confirmation bias in the outcome-based condition, which would induce a strong motivation to confirm that the chosen action does lead to the target. In other words, it is possible that having a target (e.g. "draw orange") may bias subjects towards believing they are drawing from that target even when they are not, and as a result needing less evidence to reach that conclusion (similar to the effect shown in Gesiarz, Cahill and Sharot, 2019, Plos Computational Biology). This could in turn lead to the observed patterns of results, i.e. needing more evidence against orange to switch, being less confident when switching etc... Additionally, the result that prior beliefs are stronger in the Ob condition (p.18-19) is consistent with this idea, since confirmation bias is usually associated with stronger prior beliefs.

    2. There are a few other instances where I believe the conclusions are not fully supported by the results.
      2a) One such instance is Experiment 3, which was run to test whether the temporal direction of the inference (prospective vs retrospective) could in fact explain the observed differences between condition. First, the effects on change of mind are not shown for this Experiment, only the effects on confidence - why is that? Second the confidence data presented in Figure S1 still shows some differences: lower overall confidence overall, higher confidence time-constant, and higher confidence-PSE in the prospective vs retrospective condition. Those differences are reported but then the authors still conclude that these results constitute evidence that the observed differences between Ob and Cb conditions in the main experiment are due to controllability and not to the temporal orientation. In my opinion, this conclusion is not supported by the data. One could conclude that temporality has no effect only if no difference in choice or confidence were observed.
      2b) Another instance is the claim that there is a causal role for confidence in controlling changes-of-mind (p.11). Why this is an interesting idea, I am not sure it can be fully evidenced in this task without an experimentally controlled manipulation of confidence. The reason is that there could be a common cause to both high confidence and high propensity to confirm switch decisions without the two processes actually being causally related. One such common cause could be the strength of evidence.

    3. The introduction is a little thin and would benefit from being expanded. Specifically, the mention that "information-seeking has been mostly studied under 'exploration-exploitation' dilemma is not true - while fairly recent there are many studies that have studied information-seeking in humans using other paradigms. Second and most importantly, the introduction is lacking the rationale for the proposed work - in addition to dissociating information-seeking from changes-of-mind, which is more of a methodological aim, what are the key questions that the authors trying to address, and what are the hypotheses given the current literature? Finally, the use of MEG and other physiological measures (pupillometric, cardiac patterns) is not motivated at all. The introduction should set the stage as to why collecting these data is needed given the question of interest. Similarly, the hypotheses should also set the stage for the specific analysis choices performed later, i.e. in what way does each presented analysis answer the question? This is not clear at the moment.

    4. Behaviorally, one of the main novel aspects of the study is the inclusion and analysis of confidence in addition to changes of mind. Yet, the neurophysiological analyses do not touch on confidence at all, which also contributes to making the rationale for using these neurophysiological methods less clear, as described above.

    Was this evaluation helpful?