The effect of uncertainty on prediction error in the action perception loop.
Ontology highlight
ABSTRACT:
SUBMITTER: Perrykkad K
PROVIDER: S-EPMC8085733 | biostudies-literature |
REPOSITORIES: biostudies-literature
Similar Datasets
Project description:Sensory prediction-error is vital to discriminating whether sensory inputs are caused externally or are the consequence of self-action, thereby contributing to a stable perception of the external world and building sense of agency. However, it remains unexplored whether prediction error of self-action is also used to estimate the internal body condition. To address this point, we examined whether prediction error affects the perceived intensity of muscle fatigue. Participants evaluated fatigue while maintaining repetitive finger movements. To provide prediction error, we inserted a temporal delay into online visual feedback of self-movements. The results show that the subjective rating of muscle fatigue significantly increased under the delayed visual feedback, suggesting that prediction error enhances the perception of muscle fatigue. Furthermore, we introduced visual feedback that preceded actual finger movements to test whether the temporal direction of the mismatch is crucial in estimating muscle fatigue. We found that perceived fatigue was significantly weaker with preceding visual feedback compared to normal feedback, showing that the perception of muscle fatigue is affected by the signed prediction-error. Our findings support the idea that the brain flexibly attributes prediction errors to a self-origin with keeping sense of agency, or external origin by considering contexts and error characteristics.
| S-EPMC9537327 | biostudies-literature
Project description:Human speech perception can be described as Bayesian perceptual inference but how are these Bayesian computations instantiated neurally? We used magnetoencephalographic recordings of brain responses to degraded spoken words and experimentally manipulated signal quality and prior knowledge. We first demonstrate that spectrotemporal modulations in speech are more strongly represented in neural responses than alternative speech representations (e.g. spectrogram or articulatory features). Critically, we found an interaction between speech signal quality and expectations from prior written text on the quality of neural representations; increased signal quality enhanced neural representations of speech that mismatched with prior expectations, but led to greater suppression of speech that matched prior expectations. This interaction is a unique neural signature of prediction error computations and is apparent in neural responses within 100 ms of speech input. Our findings contribute to the detailed specification of a computational model of speech perception based on predictive coding frameworks.
| S-EPMC7641582 | biostudies-literature
Project description:Human movements are prone to errors that arise from inaccuracies in both our perceptual processing and execution of motor commands. We can reduce such errors by both improving our estimates of the state of the world and through online error correction of the ongoing action. Two prominent frameworks that explain how humans solve these problems are Bayesian estimation and stochastic optimal feedback control. Here we examine the interaction between estimation and control by asking if uncertainty in estimates affects how subjects correct for errors that may arise during the movement. Unbeknownst to participants, we randomly shifted the visual feedback of their finger position as they reached to indicate the center of mass of an object. Even though participants were given ample time to compensate for this perturbation, they only fully corrected for the induced error on trials with low uncertainty about center of mass, with correction only partial in trials involving more uncertainty. The analysis of subjects' scores revealed that participants corrected for errors just enough to avoid significant decrease in their overall scores, in agreement with the minimal intervention principle of optimal feedback control. We explain this behavior with a term in the loss function that accounts for the additional effort of adjusting one's response. By suggesting that subjects' decision uncertainty, as reflected in their posterior distribution, is a major factor in determining how their sensorimotor system responds to error, our findings support theoretical models in which the decision making and control processes are fully integrated.
| S-EPMC5271325 | biostudies-literature
Project description:Judgments of the orientation of a visual line with respect to earth vertical are affected by panoramic visual cues. This is illustrated by the rod-and-frame effect (RFE), the finding that the perceived orientation of a luminous rod is biased by the orientation of a surrounding squared frame. In this study, we tested how the uncertainty of frame orientation affects the RFE by asking upright or tilted participants to psychometrically judge the orientation of a briefly flashed rod contained within either a circular frame, a squared frame, or either of two intermediate frame forms, called squircles, presented in various orientations. Results showed a cyclical modulation of frame-induced bias across the range of the square and squircular frame orientations. The magnitude of this bias increased with increasing squaredness of the frame, as if the more unequivocal the orientation cues of the frame, the larger the reliance on them for rod orientation judgments. These findings are explained with a Bayesian optimal integration model in which participants flexibly weigh visual panoramic cues, depending on their orientation reliability, and non-visual cues in the perception of vertical.
| S-EPMC8635489 | biostudies-literature
Project description:A prominent and influential hypothesis of vision suggests the existence of two separate visual systems within the brain, one creating our perception of the world and another guiding our actions within it. The induced Roelofs effect has been described as providing strong evidence for this perception/action dissociation: When a small visual target is surrounded by a large frame positioned so that the frame's center is offset from the observer's midline, the perceived location of the target is shifted in the direction opposite the frame's offset. In spite of this perceptual mislocalization, however, the observer can accurately guide movements to the target location. Thus, perception is prone to the illusion while actions seem immune. Here we demonstrate that the Roelofs illusion is caused by a frame-induced transient distortion of the observer's apparent midline. We further demonstrate that actions guided to targets within this same distorted egocentric reference frame are fully expected to be accurate, since the errors of target localization will exactly cancel the errors of motor guidance. These findings provide a mechanistic explanation for the various perceptual and motor effects of the induced Roelofs illusion without requiring the existence of separate neural systems for perception and action. Given this, the behavioral dissociation that accompanies the Roelofs effect cannot be considered evidence of a dissociation of perception and action. This indicates a general need to re-evaluate the broad class of evidence purported to support this hypothesized dissociation.
| S-EPMC524248 | biostudies-literature
Project description:Social and linguistic perceptions are linked. On one hand, talker identity affects speech perception. On the other hand, speech itself provides information about a talker's identity. Here, we propose that the same probabilistic knowledge might underlie both socially conditioned linguistic inferences and linguistically conditioned social inferences. Our computational-level approach-the ideal adapter-starts from the idea that listeners use probabilistic knowledge of covariation between social, linguistic, and acoustic cues in order to infer the most likely explanation of the speech signals they hear. As a first step toward understanding social inferences in this framework, we use a simple ideal observer model to show that it would be possible to infer aspects of a talker's identity using cue distributions based on actual speech production data. This suggests the possibility of a single formal framework for social and linguistic inferences and the interactions between them.
| S-EPMC6139081 | biostudies-literature
Project description:Affordances represent features of an object that trigger specific actions. Here we tested whether the presence and orientation of a handle on a cup could bias grasping movements towards it in conditions where subjects were explicitly told to ignore the handle. We quantified the grip aperture profile of twelve healthy participants instructed to grasp a cup from its body while it either had no handle, a handle pointing towards, or away from the grasping hand (3 'move' conditions, with large grip aperture). To ensure the smaller grip aperture afforded by the handle was implicitly processed, we interspersed trials in which participants had to grasp the cup from its handle or a handle not attached to a cup with a small grip aperture. We found that grip aperture was smaller in the presence of a handle in the 'move' conditions, independently of its orientation. Our finding, of an effect of the handle during the execution of a grasp action, extends previous evidence of such an influence measured during motor preparation using simple reaction times. It suggests that the specific action elicited by an object's attribute can affect movement performance in a sustained manner throughout movement execution.
| S-EPMC5760564 | biostudies-literature
Project description:BackgroundThe biomarker discovery field is replete with molecular signatures that have not translated into the clinic despite ostensibly promising performance in predicting disease phenotypes. One widely cited reason is lack of classification consistency, largely due to failure to maintain performance from study to study. This failure is widely attributed to variability in data collected for the same phenotype among disparate studies, due to technical factors unrelated to phenotypes (e.g., laboratory settings resulting in "batch-effects") and non-phenotype-associated biological variation in the underlying populations. These sources of variability persist in new data collection technologies.MethodsHere we quantify the impact of these combined "study-effects" on a disease signature's predictive performance by comparing two types of validation methods: ordinary randomized cross-validation (RCV), which extracts random subsets of samples for testing, and inter-study validation (ISV), which excludes an entire study for testing. Whereas RCV hardwires an assumption of training and testing on identically distributed data, this key property is lost in ISV, yielding systematic decreases in performance estimates relative to RCV. Measuring the RCV-ISV difference as a function of number of studies quantifies influence of study-effects on performance.ResultsAs a case study, we gathered publicly available gene expression data from 1,470 microarray samples of 6 lung phenotypes from 26 independent experimental studies and 769 RNA-seq samples of 2 lung phenotypes from 4 independent studies. We find that the RCV-ISV performance discrepancy is greater in phenotypes with few studies, and that the ISV performance converges toward RCV performance as data from additional studies are incorporated into classification.ConclusionsWe show that by examining how fast ISV performance approaches RCV as the number of studies is increased, one can estimate when "sufficient" diversity has been achieved for learning a molecular signature likely to translate without significant loss of accuracy to new clinical settings.
| S-EPMC4201588 | biostudies-literature
Project description:Prior work has yielded contradictory findings regarding the association between depression and the error-related negativity (ERN), an event-related potential thought to reflect individual variation in sensitivity to internal threat (i.e., errors), and the error positivity (Pe), which is thought to reflect more elaborative, conscious processing of errors. One possibility is that variation in transdiagnostic dimensional constructs related to threat processing might help explain inconsistencies in the relationship between depression and the ERN and Pe. Here, we used a large, unselected sample (N = 100) to determine whether variation in intolerance of uncertainty (IU), a transdiagnostic trait dimension that is implicated in both depression and anxiety, might moderate associations between depressive symptomatology and error processing. Results showed that greater levels of depressive symptomatology were associated with larger ΔERNs (error minus correct trials) when IU was low, but were unrelated to ΔERN when IU was high; main effects of depression and IU on ΔERN were not observed. The Pe was not associated with IU or depression. Additionally, IU, but not depression, was associated with faster response times on error and correct trials. Overall, results suggest that error processing may differ for individuals with elevated depression with versus without elevated IU. Moreover, prior failures to observe associations between depression and the ERN might stem from failure to account for related transdiagnostic constructs.
| S-EPMC7934182 | biostudies-literature
Project description:Surprise drives learning. Various neural "prediction error" signals are believed to underpin surprise-based reinforcement learning. Here, we report a surprise signal that reflects reinforcement learning but is neither un/signed reward prediction error (RPE) nor un/signed state prediction error (SPE). To exclude these alternatives, we measured surprise responses in the absence of RPE and accounted for a host of potential SPE confounds. This new surprise signal was evident in ventral striatum, primary sensory cortex, frontal poles, and amygdala. We interpret these findings via a normative model of surprise.
| S-EPMC4312927 | biostudies-literature