Not Noisy, Just Wrong: The Role of Suboptimal Inference in Behavioral Variability

Beck, Jeffrey M.; Ji, Wei; Pitkow, Xaq; Latham, Peter E.; Pouget, Alexandre

doi:10.1016/j.neuron.2012.03.016

Cited by 300 publications

(318 citation statements)

References 64 publications

Supporting

Mentioning

309

Contrasting

Order By: Relevance

“…Here, the linear Fisher information estimated the amount of information about stimulus direction available in a locally optimal linear estimator (Beck et al, 2012;Moreno-Bote et al, 2014). The linear Fisher information is inversely proportional to the square of the discrimination threshold, which is the smallest difference between two stimulus directions that can be correctly determined by an optimal decoder of the neural activity.…”

Section: Measure Of Population Informationmentioning

confidence: 99%

Cue Reliability Represented in the Shape of Tuning Curves in the Owl's Sound Localization System

2016

View full text Add to dashboard Cite

Optimal use of sensory information requires that the brain estimates the reliability of sensory cues, but the neural correlate of cue reliability relevant for behavior is not well defined. Here, we addressed this issue by examining how the reliability of spatial cue influences neuronal responses and behavior in the owl's auditory system. We show that the firing rate and spatial selectivity changed with cue reliability due to the mechanisms generating the tuning to the sound localization cue. We found that the correlated variability among neurons strongly depended on the shape of the tuning curves. Finally, we demonstrated that the change in the neurons' selectivity was necessary and sufficient for a network of stochastic neurons to predict behavior when sensory cues were corrupted with noise. This study demonstrates that the shape of tuning curves can stand alone as a coding dimension of environmental statistics.

show abstract

Section: Measure Of Population Informationmentioning

confidence: 99%

Cue Reliability Represented in the Shape of Tuning Curves in the Owl's Sound Localization System

2016

View full text Add to dashboard Cite

show abstract

“…One possibility is that the Bayesian computations are suboptimal because incorrect generative models are being used by observers. Beck, Ma, Pitkow, Latham, and Pouget (2012) suggest suboptimal inference is inevitable, especially in complex tasks such as object recognition where the full specification of the generative model (the physics of light interacting with surfaces) is impossible due to its complexity. Alternatively, there could be limitations upon the ability to learn and represent complex prior distributions (Acerbi, Vijayakumar, & Wolpert, 2014).…”

Section: Bayes and Optimalitymentioning

confidence: 99%

Bayesian accounts of covert selective attention: A tutorial review

Vincent

2015

Atten Percept Psychophys

View full text Add to dashboard Cite

Decision making and optimal observer models offer an important theoretical approach to the study of covert selective attention. While their probabilistic formulation allows quantitative comparison to human performance, the models can be complex and their insights are not always immediately apparent. Part 1 establishes the theoretical appeal of the Bayesian approach, and introduces the way in which probabilistic approaches can be applied to covert search paradigms. Part 2 presents novel formulations of Bayesian models of 4 important covert attention paradigms, illustrating optimal observer predictions over a range of experimental manipulations. Graphical model notation is used to present models in an accessible way and Supplementary Code is provided to help bridge the gap between model theory and practical implementation. Part 3 reviews a large body of empirical and modelling evidence showing that many experimental phenomena in the domain of covert selective attention are a set of by-products. These effects emerge as the result of observers conducting Bayesian inference with noisy sensory observations, prior expectations, and knowledge of the generative structure of the stimulus environment.

show abstract

“…This can give valuable insights into how coding is affected by aspects of the neural population response. However, correlations and tuning curves are not intrinsic properties of a population response that can be changed arbitrarily (Shea-Brown et al, 2008;Beck et al, 2012). To examine how the statistics of neuronal activity affect coding, it is therefore important to consider realistic networks with realistic inputs.…”

Section: Discussionmentioning

confidence: 99%

Near-Optimal Decoding of Transient Stimuli from Coupled Neuronal Subpopulations

et al. 2014

View full text Add to dashboard Cite

Coupling between sensory neurons impacts their tuning properties and correlations in their responses. How such coupling affects sensory representations and ultimately behavior remains unclear. We investigated the role of neuronal coupling during visual processing using a realistic biophysical model of the vertical system (VS) cell network in the blow fly. These neurons are thought to encode the horizontal rotation axis during rapid free-flight maneuvers. Experimental findings suggest that neurons of the VS are strongly electrically coupled, and that several downstream neurons driving motor responses to ego-rotation receive inputs primarily from a small subset of VS cells. These downstream neurons must decode information about the axis of rotation from a partial readout of the VS population response. To investigate the role of coupling, we simulated the VS response to a variety of rotating visual scenes and computed optimal Bayesian estimates from the relevant subset of VS cells. Our analysis shows that coupling leads to near-optimal estimates from a subpopulation readout. In contrast, coupling between VS cells has no impact on the quality of encoding in the response of the full population. We conclude that coupling at one level of the fly visual system allows for near-optimal decoding from partial information at the subsequent, premotor level. Thus, electrical coupling may provide a general mechanism to achieve near-optimal information transfer from neuronal subpopulations across organisms and modalities.

show abstract

Not Noisy, Just Wrong: The Role of Suboptimal Inference in Behavioral Variability

Cited by 300 publications

References 64 publications

Cue Reliability Represented in the Shape of Tuning Curves in the Owl's Sound Localization System

Cue Reliability Represented in the Shape of Tuning Curves in the Owl's Sound Localization System

Bayesian accounts of covert selective attention: A tutorial review

Near-Optimal Decoding of Transient Stimuli from Coupled Neuronal Subpopulations

Contact Info

Product

Resources

About