Behavioural data obtained with perceptual decision making experiments are typically analysed the drift-diffusion model. This parsimonious model accumulates noisy pieces of evidence towards a bound to explain accuracy and reaction times subjects. Recently, Bayesian models have been proposed how brain extracts information from input as presented in tasks. It has long known that is tightly linked such functional but precise relationship two mechanisms was never made explicit. Using model, we...

Even for simple perceptual decisions, the mechanisms that brain employs are still under debate. Although current consensus states accumulates evidence extracted from noisy sensory information, open questions remain about how this model relates to other phenomena such as flexibility in decision-dependent modulation of gain, or confidence a decision. We propose novel approach decisions made by combining two influential formalisms into new model. Specifically, we embed an attractor decision...

Reinforcement learning in the high-dimensional, continuous spaces typical robotics, remains a challenging problem. To overcome this challenge, popular approach has been to use demonstrations find an appropriate initialisation of policy attempt reduce number iterations needed solution. Here, we present alternative way incorporate prior knowledge from individual postures into learning, by extracting inherent problem structure efficient state representation. In particular, probabilistic,...

Perceptual decision making can be described as a process of accumulating evidence to bound which has been formalized within drift-diffusion models (DDMs). Recently, an equivalent Bayesian model proposed. In contrast standard DDMs, this directly links information in the stimulus process. Here, we extend further and allow inter-trial variability two parameters following extended version DDM. We derive parameter distributions for show that they lead predictions are qualitatively those made by...

There is ample evidence that the brain generates predictions help interpret sensory input. To build such capitalizes upon learned statistical regularities and associations (e.g., "A" followed by "B"; "C" appears together with "D"). The centrality of to mental activities gave rise hypothesis associative information predictive value perceived as intrinsically valuable. Such would ensure this proactively searched for, thereby promoting certainty stability in our environment. We therefore tested...

Trial-and-error learning is a universal strategy for establishing which actions are beneficial or harmful in new environments. However, stimulus-response associations solely via trial-and-error often suboptimal, as many settings dependencies among stimuli and responses can be exploited to increase efficiency. Previous studies have shown that featuring such dependencies, humans typically engage high-level cognitive processes employ advanced strategies improve their Here we analyze detail the...

Abstract Decisions in everyday life are prone to error. Standard models typically assume that errors during perceptual decisions due noise. However, it is unclear how noise the sensory input affects decision. Here we show there experimental tasks for which one can analyse exact spatio-temporal details of a dynamic and better understand variability human decisions. Using new visual tracking task novel Bayesian decision making model, found fluctuations single trials explain significant part...

Perceptual decisions not only depend on the incoming information from sensory systems but constitute a combination of current evidence and internally accumulated past encounters. Although recent emphasizes fundamental role prior knowledge for perceptual decision making, few studies have quantified relevance such priors examined their interplay with other decision-relevant factors, as stimulus properties. In present study we asked whether hysteresis, describing stability percept despite...

Adaptive motor control under continuously varying context, like the inertia parameters of a ma- nipulated object, is an active research area that lacks satisfactory solution. Here, we present and compare three novel strategies for learning context show how adding tactile sensors may ease this task. The first strategy uses only dynamics information to infer unknown parameters. It based on probabilistic generative model torques, which are linear in We demonstrate inference special case single...

In recent years, the advent of great technological advances has produced a wealth very high-dimensional data, and combining information from multiple sources is becoming increasingly important in an extending range scientific disciplines. Partial Least Squares Correlation (PLSC) frequently used method for multivariate multimodal data integration. It is, however, computationally extensive applications involving large numbers variables, as required, example, functional genomics. To handle...

In perceptual decision making the brain extracts and accumulates evidence from a stimulus over time eventually makes based on accumulated evidence. Several characteristics of this process have been observed in human electrophysiological experiments, especially an average build-up motor-related signals supposedly reflecting evidence, when averaged across trials. Another recently established approach to investigate representation is correlate within-trial fluctuations with measured signals. We...

The olfactory information that is received by the insect brain encoded in form of spatiotemporal patterns projection neurons antennal lobe. These dense and overlapping are transformed into a sparse code Kenyon cells mushroom body. Although it clear this basis for rapid categorization odors, yet unclear how computed what represents. Here we show computation can be modeled sequential firing rate using Lotka-Volterra equations Bayesian online inference. This new model understood as an...

In simple perceptual decisions the brain has to identify a stimulus based on noisy sensory samples from stimulus. Basic statistical considerations state that reliability of information, i.e., amount noise in samples, should be taken into account when decision is made. However, for making experiments it been questioned whether indeed uses confronted with unpredictable changes reliability. We here show even basic drift diffusion model, which frequently used explain experimental findings...

Die Wahrnehmung eines Reizes wird selten durch seine Eigenschaften alleine bestimmt. Sie ist vielmehr ein effizienter, dynamischer Prozess, der die eingehende sensorische Information mit bereits bestehender verknüpft. Dies kann auf Grundlage des Entscheidungsprozesses zu einem kurzzeitig vorangegangenen Stimulus erfolgen, somit Entscheidung nachfolgenden prägt. Dieser Effekt als perzeptuelle Hysterese bekannt. In dieser Studie wurde Einfluss perzeptuellen Entscheidungen bei taktiler untersucht.

Abstract Fast and accurate decisions are fundamental for adaptive behaviour. Theories of decision making posit that evidence in favour different choices is gradually accumulated until a critical value reached. It remains unclear, however, which aspects the neural code get updated during accumulation. Here we investigated whether accumulation relies on gradual increase precision representations sensory input. Healthy human volunteers discriminated global motion direction over patch moving...

Trial-and-error learning is a universal strategy for establishing which actions are beneficial or harmful in new environments. However, stimulus-response associations solely via trial-and-error often suboptimal, as many settings dependencies among stimuli and responses can be exploited to increase efficiency. Previous studies have shown that featuring such dependencies, humans typically engage high-level cognitive processes employ advanced strategies improve their Here we analyze detail the...