Neurally constrained modeling of perceptual decision making

Does allowing for changes of mind influence initial responses?

Evidence accumulation models (EAMs) have become the dominant theoretical framework for rapid decision-making, and while many theoretically distinct variants exist, comparisons have proved challenging due to strong mimicry in their predictions about choice response time data. One solution to reduce mimicry is constraining these models with double responses, which are a second response that is made after the initial response. However, instructing participants that they are allowed to change their mind could influence their strategy for initial responding, meaning that explicit double responding paradigms may not generalise to standard paradigms. Here, we provide a validation of explicit double responding paradigms, by assessing whether participants' initial decisions - as measured by diffusion model parameters - differ based on whether or not they were instructed that they could change their response after their initial response. Across three experiments, our results consistently indicate that allowing for changes of mind does not influence initial responses, with Bayesian analyses providing at least moderate evidence in favour of the null in all cases. Our findings suggest that explicit double responding paradigms should generalise to standard paradigms, validating the use of explicit double responding in future rapid decision-making studies.

Dynamic computational phenotyping of human cognition

Computational phenotyping has emerged as a powerful tool for characterizing individual variability across a variety of cognitive domains. An individual's computational phenotype is defined as a set of mechanistically interpretable parameters obtained from fitting computational models to behavioural data. However, the interpretation of these parameters hinges critically on their psychometric properties, which are rarely studied. To identify the sources governing the temporal variability of the computational phenotype, we carried out a 12-week longitudinal study using a battery of seven tasks that measure aspects of human learning, memory, perception and decision making. To examine the influence of state effects, each week, participants provided reports tracking their mood, habits and daily activities. We developed a dynamic computational phenotyping framework, which allowed us to tease apart the time-varying effects of practice and internal states such as affective valence and arousal. Our results show that many phenotype dimensions covary with practice and affective factors, indicating that what appears to be unreliability may reflect previously unmeasured structure. These results support a fundamentally dynamic understanding of cognitive variability within an individual.

Neural mechanisms for executive control of speed-accuracy trade-off

The medial frontal cortex (MFC) plays an important but disputed role in speed-accuracy trade-off (SAT). In samples of neural spiking in the supplementary eye field (SEF) in the MFC simultaneous with the visuomotor frontal eye field and superior colliculus in macaques performing a visual search with instructed SAT, during accuracy emphasis, most SEF neurons discharge less from before stimulus presentation until response generation. Discharge rates adjust immediately and simultaneously across structures upon SAT cue changes. SEF neurons signal choice errors with stronger and earlier activity during accuracy emphasis. Other neurons signal timing errors, covarying with adjusting response time. Spike correlations between neurons in the SEF and visuomotor areas did not appear, disappear, or change sign across SAT conditions or trial outcomes. These results clarify findings with noninvasive measures, complement previous neurophysiological findings, and endorse the role of the MFC as a critic for the actor instantiated in visuomotor structures.

Distributed Coding of Evidence Accumulation across the Mouse Brain Using Microcircuits with a Diversity of Timescales

The gradual accumulation of noisy evidence for or against options is the main step in the perceptual decision-making process. Using brain-wide electrophysiological recording in mice (Steinmetz et al., 2019), we examined neural correlates of evidence accumulation across brain areas. We demonstrated that the neurons with drift-diffusion model (DDM)-like firing rate activity (i.e., evidence-sensitive ramping firing rate) were distributed across the brain. Exploring the underlying neural mechanism of evidence accumulation for the DDM-like neurons revealed different accumulation mechanisms (i.e., single and race) both within and across the brain areas. Our findings support the hypothesis that evidence accumulation is happening through multiple integration mechanisms in the brain. We further explored the timescale of the integration process in the single and race accumulator models. The results demonstrated that the accumulator microcircuits within each brain area had distinct properties in terms of their integration timescale, which were organized hierarchically across the brain. These findings support the existence of evidence accumulation over multiple timescales. Besides the variability of integration timescale across the brain, a heterogeneity of timescales was observed within each brain area as well. We demonstrated that this variability reflected the diversity of microcircuit parameters, such that accumulators with longer integration timescales had higher recurrent excitation strength.

Feedforward attentional selection in sensory cortex

Salient objects grab attention because they stand out from their surroundings. Whether this phenomenon is accomplished by bottom-up sensory processing or requires top-down guidance is debated. We tested these alternative hypotheses by measuring how early and in which cortical layer(s) neural spiking distinguished a target from a distractor. We measured synaptic and spiking activity across cortical columns in mid-level area V4 of male macaque monkeys performing visual search for a color singleton. A neural signature of attentional capture was observed in the earliest response in the input layer 4. The magnitude of this response predicted response time and accuracy. Errant behavior followed errant selection. Because this response preceded top-down influences and arose in the cortical layer not targeted by top-down connections, these findings demonstrate that feedforward activation of sensory cortex can underlie attentional priority.

Balancing true and false detection of intermittent sensory targets by adjusting the inputs to the evidence accumulation process

Decisions about noisy stimuli are widely understood to be made by accumulating evidence up to a decision bound that can be adjusted according to task demands. However, relatively little is known about how such mechanisms operate in continuous monitoring contexts requiring intermittent target detection. Here, we examined neural decision processes underlying detection of 1 s coherence targets within continuous random dot motion, and how they are adjusted across contexts with weak, strong, or randomly mixed weak/strong targets. Our prediction was that decision bounds would be set lower when weak targets are more prevalent. Behavioural hit and false alarm rate patterns were consistent with this, and were well captured by a bound-adjustable leaky accumulator model. However, beta-band EEG signatures of motor preparation contradicted this, instead indicating lower bounds in the strong-target context. We thus tested two alternative models in which decision-bound dynamics were constrained directly by beta measurements, respectively, featuring leaky accumulation with adjustable leak, and non-leaky accumulation of evidence referenced to an adjustable sensory-level criterion. We found that the latter model best explained both behaviour and neural dynamics, highlighting novel means of decision policy regulation and the value of neurally informed modelling.

Secondary motor integration as a final arbiter in sensorimotor decision-making

Sensorimotor decision-making is believed to involve a process of accumulating sensory evidence over time. While current theories posit a single accumulation process prior to planning an overt motor response, here, we propose an active role of motor processes in decision formation via a secondary leaky motor accumulation stage. The motor leak adapts the "memory" with which this secondary accumulator reintegrates the primary accumulated sensory evidence, thus adjusting the temporal smoothing in the motor evidence and, correspondingly, the lag between the primary and motor accumulators. We compare this framework against different single accumulator variants using formal model comparison, fitting choice, and response times in a task where human observers made categorical decisions about a noisy sequence of images, under different speed-accuracy trade-off instructions. We show that, rather than boundary adjustments (controlling the amount of evidence accumulated for decision commitment), adjustment of the leak in the secondary motor accumulator provides the better description of behavior across conditions. Importantly, we derive neural correlates of these 2 integration processes from electroencephalography data recorded during the same task and show that these neural correlates adhere to the neural response profiles predicted by the model. This framework thus provides a neurobiologically plausible description of sensorimotor decision-making that captures emerging evidence of the active role of motor processes in choice behavior.

Neural Mechanisms That Make Perceptual Decisions Flexible

Neural mechanisms of perceptual decision making have been extensively studied in experimental settings that mimic stable environments with repeating stimuli, fixed rules, and payoffs. In contrast, we live in an ever-changing environment and have varying goals and behavioral demands. To accommodate variability, our brain flexibly adjusts decision-making processes depending on context. Here, we review a growing body of research that explores the neural mechanisms underlying this flexibility. We highlight diverse forms of context dependency in decision making implemented through a variety of neural computations. Context-dependent neural activity is observed in a distributed network of brain structures, including posterior parietal, sensory, motor, and subcortical regions, as well as the prefrontal areas classically implicated in cognitive control. We propose that investigating the distributed network underlying flexible decisions is key to advancing our understanding and discuss a path forward for experimental and theoretical investigations.

Strategically managing learning during perceptual decision making

Making optimal decisions in the face of noise requires balancing short-term speed and accuracy. But a theory of optimality should account for the fact that short-term speed can influence long-term accuracy through learning. Here, we demonstrate that long-term learning is an important dynamical dimension of the speed-accuracy trade-off. We study learning trajectories in rats and formally characterize these dynamics in a theory expressed as both a recurrent neural network and an analytical extension of the drift-diffusion model that learns over time. The model reveals that choosing suboptimal response times to learn faster sacrifices immediate reward, but can lead to greater total reward. We empirically verify predictions of the theory, including a relationship between stimulus exposure and learning speed, and a modulation of reaction time by future learning prospects. We find that rats' strategies approximately maximize total reward over the full learning epoch, suggesting cognitive control over the learning process.

Efficiency in sequential testing: Comparing the sequential probability ratio test and the sequential Bayes factor test

In a sequential hypothesis test, the analyst checks at multiple steps during data collection whether sufficient evidence has accrued to make a decision about the tested hypotheses. As soon as sufficient information has been obtained, data collection is terminated. Here, we compare two sequential hypothesis testing procedures that have recently been proposed for use in psychological research: Sequential Probability Ratio Test (SPRT; Psychological Methods, 25(2), 206-226, 2020) and the Sequential Bayes Factor Test (SBFT; Psychological Methods, 22(2), 322-339, 2017). We show that although the two methods have different philosophical roots, they share many similarities and can even be mathematically regarded as two instances of an overarching hypothesis testing framework. We demonstrate that the two methods use the same mechanisms for evidence monitoring and error control, and that differences in efficiency between the methods depend on the exact specification of the statistical models involved, as well as on the population truth. Our simulations indicate that when deciding on a sequential design within a unified sequential testing framework, researchers need to balance the needs of test efficiency, robustness against model misspecification, and appropriate uncertainty quantification. We provide guidance for navigating these design decisions based on individual preferences and simulation-based design analyses.

Frontal eye fields in macaque monkeys: prefrontal and premotor contributions to visually guided saccades

Neuronal spiking was sampled from the frontal eye field (FEF) and from the rostral part of area 6 that reaches to the superior limb of the arcuate sulcus, dorsal to the arcuate spur when present (F2vr) in macaque monkeys performing memory-guided saccades and visually guided saccades for visual search. Neuronal spiking modulation in F2vr resembled that in FEF in many but not all respects. A new consensus clustering algorithm of neuronal modulation patterns revealed that F2vr and FEF contain a greater variety of modulation patterns than previously reported. The areas differ in the proportions of visuomotor neuron types, the proportions of neurons discriminating a target from distractors during visual search, and the consistency of modulation patterns across tasks. However, between F2vr and FEF we found no difference in the magnitude of delay period activity, the timing of the peak discharge rate relative to saccades, or the time of search target selection. The observed similarities and differences between the 2 cortical regions contribute to other work establishing the organization of eye fields in the frontal lobe and may help explain why FEF in monkeys is identified within granular prefrontal area 8 but in humans is identified within agranular premotor area 6.

Relating a Spiking Neural Network Model and the Diffusion Model of Decision-Making

Many models of decision making assume accumulation of evidence to threshold as a core mechanism to predict response probabilities and response times. A spiking neural network model (Wang, 2002) instantiates these mechanisms at the level of biophysically-plausible pools of neurons with excitatory and inhibitory connections, and has numerous model parameters tuned by physiological measures. The diffusion model (Ratcliff, 1978) is a cognitive model that can be fitted to a range of behaviors and conditions. We investigated how parameters of the cognitive-level diffusion model relate to the parameters of a neural-level spiking model. In each simulated "experiment", we generated "data" from the spiking neural network by factorially combining a manipulation of choice difficulty (via the input to the spiking model) and a manipulation of one of the core parameters of the spiking model. We then fitted the diffusion model to these simulated data to observe how manipulation of each core spiking model parameter mapped on to fitted drift rate, response threshold, and non-decision time. Manipulations of parameters in the spiking model related to input sensitivity, threshold, and stimulus processing time mapped on to their conceptual analogues in the diffusion model, namely drift rate, threshold, and non-decision time. Manipulations of parameters in the spiking model with no direct analogue to the diffusion model, non-stimulus-specific background input, strength of recurrent excitation, and receptor conductances, mapped on to threshold in the diffusion model. We discuss implications of these results for interpretations of fits of the diffusion model to behavioral data.

Reassessing hierarchical correspondences between brain and deep networks through direct interface

Functional correspondences between deep convolutional neural networks (DCNNs) and the mammalian visual system support a hierarchical account in which successive stages of processing contain ever higher-level information. However, these correspondences between brain and model activity involve shared, not task-relevant, variance. We propose a stricter account of correspondence: If a DCNN layer corresponds to a brain region, then replacing model activity with brain activity should successfully drive the DCNN's object recognition decision. Using this approach on three datasets, we found that all regions along the ventral visual stream best corresponded with later model layers, indicating that all stages of processing contained higher-level information about object category. Time course analyses suggest that long-range recurrent connections transmit object class information from late to early visual areas.

Developing control-theoretic objectives for large-scale brain dynamics and cognitive enhancement

The development of technologies for brain stimulation provides a means for scientists and clinicians to directly actuate the brain and nervous system. Brain stimulation has shown intriguing potential in terms of modifying particular symptom clusters in patients and behavioral characteristics of subjects. The stage is thus set for optimization of these techniques and the pursuit of more nuanced stimulation objectives, including the modification of complex cognitive functions such as memory and attention. Control theory and engineering will play a key role in the development of these methods, guiding computational and algorithmic strategies for stimulation. In particular, realizing this goal will require new development of frameworks that allow for controlling not only brain activity, but also latent dynamics that underlie neural computation and information processing. In the current opinion, we review recent progress in brain stimulation and outline challenges and potential research pathways associated with exogenous control of cognitive function.

Behavior in a visual search task with moving dot stimuli

Understanding the neuronal mechanisms underlying the processing of visual attention requires a well-designed behavioral task that allows investigators to clearly describe the behavioral effects of attention. Here, we introduce a behavioral paradigm in which one, two or four moving dot stimuli are used in a visual search paradigm that includes two additional attentional conditions. Two animals were trained to make a saccade to a target (a dot patch with net rightward motion) and hold central fixation if no target was present. To implement covert visual attention, we included trials in which a 100% valid spatial cue appeared and trials in which the color of the fixation point indicated, with 100% validity, which of four colored dot patches the animals should attend to. We analyzed the behavior in terms of reaction times and signal detection theory metrics d-prime (representing sensitivity) and criteria. In both animals, we found that reaction times were greater for larger set-sizes and that the introduction of an attentional cue reduced the reaction times substantially. We also found that both animals showed increases in criteria, but no change in sensitivity, as set-size increased and the attentional cues led to an increase in sensitivity, with only some change in criteria. Our results illustrate how the animals perform this task and imply that both animals chose similar strategies. Importantly, this will allow future neurophysiological studies to probe not only the effects of attention, but will give the possibility of seeing whether different neural mechanisms drive changes in criteria and d-prime.

Temporal dynamics of decision making: A synthesis of computational and neurophysiological approaches

As interest in the temporal dynamics of decision-making has grown, researchers have increasingly turned to computational approaches such as the drift diffusion model (DDM) to identify how cognitive processes unfold during choice. At the same time, technological advances in noninvasive neurophysiological methods such as electroencephalography and magnetoencephalography now allow researchers to map the neural time course of decision making with millisecond precision. Combining these approaches can potentially yield important new insights into how choices emerge over time. Here we review recent research on the computational and neurophysiological correlates of perceptual and value-based decision making, from DDM parameters to scalp potentials and oscillatory neural activity. Starting with motor response preparation, the most well-understood aspect of the decision process, we discuss evidence that urgency signals and shifts in baseline activation, rather than shifts in the physiological value of the choice-triggering response threshold, are responsible for adjusting response times under speeded choice scenarios. Research on the neural correlates of starting point bias suggests that prestimulus activity can predict biases in motor choice behavior. Finally, studies examining the time dynamics of evidence construction and evidence accumulation have identified signals at frontocentral and centroparietal electrodes associated respectively with these processes, emerging 300-500 ms after stimulus onset. These findings can inform psychological theories of decision-making, providing empirical support for attribute weighting in value-based choice while suggesting theoretical alternatives to dual-process accounts. Further research combining computational and neurophysiological approaches holds promise for providing greater insight into the moment-by-moment evolution of the decision process. This article is categorized under: Psychology > Reasoning and Decision Making Neuroscience > Cognition Economics > Individual Decision-Making.

Population temporal structure supplements the rate code during sensorimotor transformations

Sensorimotor transformations are mediated by premotor brain networks where individual neurons represent sensory, cognitive, and movement-related information. Such multiplexing poses a conundrum-how does a decoder know precisely when to initiate a movement if its inputs are active at times when a movement is not desired (e.g., in response to sensory stimulation)? Here, we propose a novel hypothesis: movement is triggered not only by an increase in firing rate but, critically, also by a reliable temporal pattern in the population response. Laminar recordings in the macaque superior colliculus (SC), a midbrain hub of orienting control, and pseudo-population analyses in SC and cortical frontal eye fields (FEFs) corroborated this hypothesis. Specifically, using a measure that captures the fidelity of the population code-here called temporal stability-we show that the temporal structure fluctuates during the visual response but becomes increasingly stable during the movement command. Importantly, we used spatiotemporally patterned microstimulation to causally test the contribution of population temporal stability in gating movement initiation and found that stable stimulation patterns were more likely to evoke a movement. Finally, a spiking neuron model was able to discriminate between stable and unstable input patterns, providing a putative biophysical mechanism for decoding temporal structure. These findings offer new insights into the long-standing debate on motor preparation and generation by situating the movement gating signal in temporal features of activity in shared neural substrates, and they highlight the importance of short-term population history in neuronal communication and behavior.

Different Forms of Variability Could Explain a Difference Between Human and Rat Decision Making

When observers make rapid, difficult perceptual decisions, their response time is highly variable from trial to trial. In a visual motion discrimination task, it has been reported that human accuracy declines with increasing response time, whereas rat accuracy increases with response time. This is of interest because different mathematical theories of decision-making differ in their predictions regarding the correlation of accuracy with response time. On the premise that perceptual decision-making mechanisms are likely to be conserved among mammals, we seek to unify the rodent and primate results in a common theoretical framework. We show that a bounded drift diffusion model (DDM) can explain both effects with variable parameters: trial-to-trial variability in the starting point of the diffusion process produces the pattern typically observed in rats, whereas variability in the drift rate produces the pattern typically observed in humans. We further show that the same effects can be produced by deterministic biases, even in the absence of parameter stochasticity or parameter change within a trial.

Transient neuronal suppression for exploitation of new sensory evidence

In noisy but stationary environments, decisions should be based on the temporal integration of sequentially sampled evidence. This strategy has been supported by many behavioral studies and is qualitatively consistent with neural activity in multiple brain areas. By contrast, decision-making in the face of non-stationary sensory evidence remains poorly understood. Here, we trained monkeys to identify and respond via saccade to the dominant color of a dynamically refreshed bicolor patch that becomes informative after a variable delay. Animals’ behavioral responses were briefly suppressed after evidence changes, and many neurons in the frontal eye field displayed a corresponding dip in activity at this time, similar to that frequently observed after stimulus onset but sensitive to stimulus strength. Generalized drift-diffusion models revealed consistency of behavior and neural activity with brief suppression of motor output, but not with pausing or resetting of evidence accumulation. These results suggest that momentary arrest of motor preparation is important for dynamic perceptual decision making.

While evidence is constantly changing during real-world decisions, little is known about how the brain deals with such changes. Here, the authors show that the brain strategically suppresses motor output via the frontal eye fields in response to stimulus changes.

Neural Substrates of the Drift-Diffusion Model in Brain Disorders

Many studies on the drift-diffusion model (DDM) explain decision-making based on a unified analysis of both accuracy and response times. This review provides an in-depth account of the recent advances in DDM research which ground different DDM parameters on several brain areas, including the cortex and basal ganglia. Furthermore, we discuss the changes in DDM parameters due to structural and functional impairments in several clinical disorders, including Parkinson's disease, Attention Deficit Hyperactivity Disorder (ADHD), Autism Spectrum Disorders, Obsessive-Compulsive Disorder (OCD), and schizophrenia. This review thus uses DDM to provide a theoretical understanding of different brain disorders.

Switching costs in stochastic environments drive the emergence of matching behaviour in animal decision-making through the promotion of reward learning strategies

A principle of choice in animal decision-making named probability matching (PM) has long been detected in animals, and can arise from different decision-making strategies. Little is known about how environmental stochasticity may influence the switching time of these different decision-making strategies. Here we address this problem using a combination of behavioral and theoretical approaches, and show, that although a simple Win-Stay-Loss-Shift (WSLS) strategy can generate PM in binary-choice tasks theoretically, budgerigars (Melopsittacus undulates) actually apply a range of sub-tactics more often when they are expected to make more accurate decisions. Surprisingly, budgerigars did not get more rewards than would be predicted when adopting a WSLS strategy, and their decisions also exhibited PM. Instead, budgerigars followed a learning strategy based on reward history, which potentially benefits individuals indirectly from paying lower switching costs. Furthermore, our data suggest that more stochastic environments may promote reward learning through significantly less switching. We suggest that switching costs driven by the stochasticity of an environmental niche can potentially represent an important selection pressure associated with decision-making that may play a key role in driving the evolution of complex cognition in animals.

Representational geometry of perceptual decisions in the monkey parietal cortex

Lateral intraparietal (LIP) neurons represent formation of perceptual decisions involving eye movements. In circuit models for these decisions, neural ensembles that encode actions compete to form decisions. Consequently, representation and readout of the decision variables (DVs) are implemented similarly for decisions with identical competing actions, irrespective of input and task context differences. Further, DVs are encoded as partially potentiated action plans through balance of activity of action-selective ensembles. Here, we test those core principles. We show that in a novel face-discrimination task, LIP firing rates decrease with supporting evidence, contrary to conventional motion-discrimination tasks. These opposite response patterns arise from similar mechanisms in which decisions form along curved population-response manifolds misaligned with action representations. These manifolds rotate in state space based on context, indicating distinct optimal readouts for different tasks. We show similar manifolds in lateral and medial prefrontal cortices, suggesting similar representational geometry across decision-making circuits.

Accumulation of continuously time-varying sensory evidence constrains neural and behavioral responses in human collision threat detection

Evidence accumulation models provide a dominant account of human decision-making, and have been particularly successful at explaining behavioral and neural data in laboratory paradigms using abstract, stationary stimuli. It has been proposed, but with limited in-depth investigation so far, that similar decision-making mechanisms are involved in tasks of a more embodied nature, such as movement and locomotion, by directly accumulating externally measurable sensory quantities of which the precise, typically continuously time-varying, magnitudes are important for successful behavior. Here, we leverage collision threat detection as a task which is ecologically relevant in this sense, but which can also be rigorously observed and modelled in a laboratory setting. Conventionally, it is assumed that humans are limited in this task by a perceptual threshold on the optical expansion rate-the visual looming-of the obstacle. Using concurrent recordings of EEG and behavioral responses, we disprove this conventional assumption, and instead provide strong evidence that humans detect collision threats by accumulating the continuously time-varying visual looming signal. Generalizing existing accumulator model assumptions from stationary to time-varying sensory evidence, we show that our model accounts for previously unexplained empirical observations and full distributions of detection response. We replicate a pre-response centroparietal positivity (CPP) in scalp potentials, which has previously been found to correlate with accumulated decision evidence. In contrast with these existing findings, we show that our model is capable of predicting the onset of the CPP signature rather than its buildup, suggesting that neural evidence accumulation is implemented differently, possibly in distinct brain regions, in collision detection compared to previously studied paradigms.

Urgent Decision Making: Resolving Visuomotor Interactions at High Temporal Resolution

Measuring when exactly perceptual decisions are made is crucial for defining how the activation of specific neurons contributes to behavior. However, in traditional, nonurgent visuomotor tasks, the uncertainty of this temporal measurement is very large. This is a problem not only for delimiting the capacity of perception, but also for correctly interpreting the functional roles ascribed to choice-related neuronal responses. In this article, we review psychophysical, neurophysiological, and modeling work based on urgent visuomotor tasks in which this temporal uncertainty can be effectively overcome. The cornerstone of this work is a novel behavioral metric that describes the evolution of the subject's perceptual judgment moment by moment, allowing us to resolve numerous perceptual events that unfold within a few tens of milliseconds. In this framework, the neural distinction between perceptual evaluation and motor selection processes becomes particularly clear, as the conclusion of one is not contingent on that of the other. Expected final online publication date for the Annual Review of Vision Science, Volume 7 is September 2021. Please see http://www.annualreviews.org/page/journal/pubdates for revised estimates.

Distractors Selectively Modulate Electrophysiological Markers of Perceptual Decisions

Current models of perceptual decision-making assume that choices are made after evidence in favor of an alternative accumulates to a given threshold. This process has recently been revealed in human EEG recordings, but an unresolved issue is how these neural mechanisms are modulated by competing, yet task-irrelevant, stimuli. In this study, we tested 20 healthy participants on a motion direction discrimination task. Participants monitored two patches of random dot motion simultaneously presented on either side of fixation for periodic changes in an upward or downward motion, which could occur equiprobably in either patch. On a random 50% of trials, these periods of coherent vertical motion were accompanied by simultaneous task-irrelevant, horizontal motion in the contralateral patch. Our data showed that these distractors selectively increased the amplitude of early target selection responses over scalp sites contralateral to the distractor stimulus, without impacting on responses ipsilateral to the distractor. Importantly, this modulation mediated a decrement in the subsequent buildup rate of a neural signature of evidence accumulation and accounted for a slowing of RTs. These data offer new insights into the functional interactions between target selection and evidence accumulation signals, and their susceptibility to task-irrelevant distractors. More broadly, these data neurally inform future models of perceptual decision-making by highlighting the influence of early processing of competing stimuli on the accumulation of perceptual evidence.

What Is the Readiness Potential?

The readiness potential (RP), a slow buildup of electrical potential recorded at the scalp using electroencephalography, has been associated with neural activity involved in movement preparation. It became famous thanks to Benjamin Libet (Brain 1983;106:623-642), who used the time difference between the RP and self-reported time of conscious intention to move to argue that we lack free will. The RP's informativeness about self-generated action and derivatively about free will has prompted continued research on this neural phenomenon. Here, we argue that recent advances in our understanding of the RP, including computational modeling of the phenomenon, call for a reassessment of its relevance for understanding volition and the philosophical problem of free will.

Reward and fictive prediction error signals in ventral striatum: asymmetry between factual and counterfactual processing

Reward prediction error, the difference between the expected and obtained reward, is known to act as a reinforcement learning neural signal. In the current study, we propose a model fitting approach that combines behavioral and neural data to fit computational models of reinforcement learning. Briefly, we penalized subject-specific fitted parameters that moved away too far from the group median, except when that deviation led to an improvement in the model's fit to neural responses. By means of a probabilistic monetary learning task and fMRI, we compared our approach with standard model fitting methods. Q-learning outperformed actor-critic at both behavioral and neural level, although the inclusion of neuroimaging data into model fitting improved the fit of actor-critic models. We observed both action-value and state-value prediction error signals in the striatum, while standard model fitting approaches failed to capture state-value signals. Finally, left ventral striatum correlated with reward prediction error while right ventral striatum with fictive prediction error, suggesting a functional hemispheric asymmetry regarding prediction-error driven learning.

The unknown but knowable relationship between Presaccadic Accumulation of activity and Saccade initiation

The goal of this short review is to call attention to a yawning gap of knowledge that separates two processes essential for saccade production. On the one hand, knowledge about the saccade generation circuitry within the brainstem is detailed and precise - push-pull interactions between gaze-shifting and gaze-holding processes control the time of saccade initiation, which begins when omnipause neurons are inhibited and brainstem burst neurons are excited. On the other hand, knowledge about the cortical and subcortical premotor circuitry accomplishing saccade initiation has crystalized around the concept of stochastic accumulation - the accumulating activity of saccade neurons reaching a fixed value triggers a saccade. Here is the gap: we do not know how the reaching of a threshold by premotor neurons causes the critical pause and burst of brainstem neurons that initiates saccades. Why this problem matters and how it can be addressed will be discussed. Closing the gap would unify two rich but curiously disconnected empirical and theoretical domains.

Sequential sampling models without random between-trial variability: the racing diffusion model of speeded decision making

Most current sequential sampling models have random between-trial variability in their parameters. These sources of variability make the models more complex in order to fit response time data, do not provide any further explanation to how the data were generated, and have recently been criticised for allowing infinite flexibility in the models. To explore and test the need of between-trial variability parameters we develop a simple sequential sampling model of N-choice speeded decision making: the racing diffusion model. The model makes speeded decisions from a race of evidence accumulators that integrate information in a noisy fashion within a trial. The racing diffusion does not assume that any evidence accumulation process varies between trial, and so, the model provides alternative explanations of key response time phenomena, such as fast and slow error response times relative to correct response times. Overall, our paper gives good reason to rethink including between-trial variability parameters in sequential sampling models.

Inhibition of saccade initiation improves saccade accuracy: The role of local and remote visual distractors in the control of saccadic eye movements

When a distractor appears close to the target location, saccades are less accurate. However, the presence of a further distractor, remote from those stimuli, increases the saccade response latency and improves accuracy. Explanations for this are either that the second, remote distractor impacts directly on target selection processes or that the remote distractor merely impairs the ability to initiate a saccade and changes the time at which unaffected target selection processes are accessed. In order to tease these two explanations apart, here we examine the relationship between latency and accuracy of saccades to a target and close distractor pair while a remote distractor appears at variable distance. Accuracy improvements are found to follow a similar pattern, regardless of the presence of the remote distractor, which suggests that the effect of the remote distractor is not the result of a direct impact on the target selection process. Our findings support the proposal that a remote distractor impairs the ability to initiate a saccade, meaning the competition between target and close distractor is accessed at a later time, thus resulting in more accurate saccades.

Neural mechanism of priming in visual search

Selective attention affords scrutinizing items in our environment. However, attentional selection changes over time and across space. Empirically, repetition of visual search conditions changes attentional processing. Priming of pop-out is a vivid example. Repeatedly searching for the same pop-out search feature is accomplished with faster response times and fewer errors. We review the psychophysical background of priming of pop-out, focusing on the hypothesis that it arises through changes in visual selective attention. We also describe research done with macaque monkeys to understand the neural mechanisms supporting visual selective attention and priming of pop-out, and survey research on priming of pop-out using noninvasive brain measures with humans. We conclude by hypothesizing three alternative neural mechanisms and highlighting open questions.

A new model of decision processing in instrumental learning tasks

Learning and decision-making are interactive processes, yet cognitive modeling of error-driven learning and decision-making have largely evolved separately. Recently, evidence accumulation models (EAMs) of decision-making and reinforcement learning (RL) models of error-driven learning have been combined into joint RL-EAMs that can in principle address these interactions. However, we show that the most commonly used combination, based on the diffusion decision model (DDM) for binary choice, consistently fails to capture crucial aspects of response times observed during reinforcement learning. We propose a new RL-EAM based on an advantage racing diffusion (ARD) framework for choices among two or more options that not only addresses this problem but captures stimulus difficulty, speed-accuracy trade-off, and stimulus-response-mapping reversal effects. The RL-ARD avoids fundamental limitations imposed by the DDM on addressing effects of absolute values of choices, as well as extensions beyond binary choice, and provides a computationally tractable basis for wider applications.

Neurocomputational mechanisms of prior-informed perceptual decision-making in humans

To interact successfully with diverse sensory environments, we must adapt our decision processes to account for time constraints and prior probabilities. The full set of decision-process parameters that undergo such flexible adaptation has proven to be difficult to establish using simplified models that are based on behaviour alone. Here, we utilize well-characterized human neurophysiological signatures of decision formation to construct and constrain a build-to-threshold decision model with multiple build-up (evidence accumulation and urgency) and delay components (pre- and post-decisional). The model indicates that all of these components were adapted in distinct ways and, in several instances, fundamentally differ from the conclusions of conventional diffusion modelling. The neurally informed model outcomes were corroborated by independent neural decision signal observations that were not used in the model's construction. These findings highlight the breadth of decision-process parameters that are amenable to strategic adjustment and the value in leveraging neurophysiological measurements to quantify these adjustments.

Frontal eye field and caudate neurons make different contributions to reward-biased perceptual decisions

Many decisions require trade-offs between sensory evidence and internal preferences. Potential neural substrates include the frontal eye field (FEF) and caudate nucleus, but their distinct roles are not understood. Previously we showed that monkeys’ decisions on a direction-discrimination task with asymmetric rewards reflected a biased accumulate-to-bound decision process (Fan et al., 2018) that was affected by caudate microstimulation (Doi et al., 2020). Here we compared single-neuron activity in FEF and caudate to each other and to accumulate-to-bound model predictions derived from behavior. Task-dependent neural modulations were similar in both regions. However, choice-selective neurons in FEF, but not caudate, encoded behaviorally derived biases in the accumulation process. Baseline activity in both regions was sensitive to reward context, but this sensitivity was not reliably associated with behavioral biases. These results imply distinct contributions of FEF and caudate neurons to reward-biased decision-making and put experimental constraints on the neural implementation of accumulation-to-bound-like computations.

Robust parallel decision-making in neural circuits with nonlinear inhibition

Animals frequently need to choose the best alternative from a set of possibilities, whether it is which direction to swim in or which food source to favor. How long should a network of neurons take to choose the best of N options? Theoretical results suggest that the optimal time grows as log(N), if the values of each option are imperfectly perceived. However, standard self-terminating neural network models of decision-making cannot achieve this optimal behavior. We show how using certain additional nonlinear response properties in neurons, which are ignored in standard models, results in a decision-making architecture that both achieves the optimal scaling of decision time and accounts for multiple experimentally observed features of neural decision-making.

An elemental computation in the brain is to identify the best in a set of options and report its value. It is required for inference, decision-making, optimization, action selection, consensus, and foraging. Neural computing is considered powerful because of its parallelism; however, it is unclear whether neurons can perform this max-finding operation in a way that improves upon the prohibitively slow optimal serial max-finding computation (which takes ∼N⁡log(N) time for N noisy candidate options) by a factor of N, the benchmark for parallel computation. Biologically plausible architectures for this task are winner-take-all (WTA) networks, where individual neurons inhibit each other so only those with the largest input remain active. We show that conventional WTA networks fail the parallelism benchmark and, worse, in the presence of noise, altogether fail to produce a winner when N is large. We introduce the nWTA network, in which neurons are equipped with a second nonlinearity that prevents weakly active neurons from contributing inhibition. Without parameter fine-tuning or rescaling as N varies, the nWTA network achieves the parallelism benchmark. The network reproduces experimentally observed phenomena like Hick’s law without needing an additional readout stage or adaptive N-dependent thresholds. Our work bridges scales by linking cellular nonlinearities to circuit-level decision-making, establishes that distributed computation saturating the parallelism benchmark is possible in networks of noisy, finite-memory neurons, and shows that Hick’s law may be a symptom of near-optimal parallel decision-making with noisy input.

Prefrontal cortex exhibits multidimensional dynamic encoding during decision-making

Recent work has suggested that prefrontal cortex (PFC) plays a key role in context-dependent perceptual decision-making. Here we address that role using a new method for identifying task-relevant dimensions of neural population activity. Specifically, we show that PFC has a multi-dimensional code for context, decisions, and both relevant and irrelevant sensory information. Moreover, these representations evolve in time, with an early linear accumulation phase followed by a phase with rotational dynamics. We identify the dimensions of neural activity associated with these phases, and show that they do not arise from distinct populations, but of a single population with broad tuning characteristics. Finally, we use model-based decoding to show that the transition from linear to rotational dynamics coincides with a plateau in decoding accuracy, revealing that rotational dynamics in PFC preserve sensory choice information for the duration of the stimulus integration period.

A flexible framework for simulating and fitting generalized drift-diffusion models

The drift-diffusion model (DDM) is an important decision-making model in cognitive neuroscience. However, innovations in model form have been limited by methodological challenges. Here, we introduce the generalized drift-diffusion model (GDDM) framework for building and fitting DDM extensions, and provide a software package which implements the framework. The GDDM framework augments traditional DDM parameters through arbitrary user-defined functions. Models are solved numerically by directly solving the Fokker-Planck equation using efficient numerical methods, yielding a 100-fold or greater speedup over standard methodology. This speed allows GDDMs to be fit to data using maximum likelihood on the full response time (RT) distribution. We demonstrate fitting of GDDMs within our framework to both animal and human datasets from perceptual decision-making tasks, with better accuracy and fewer parameters than several DDMs implemented using the latest methodology, to test hypothesized decision-making mechanisms. Overall, our framework will allow for decision-making model innovation and novel experimental designs.

Cognitive control and automatic interference in mind and brain: A unified model of saccadic inhibition and countermanding

Countermanding behavior has long been seen as a cornerstone of executive control-the human ability to selectively inhibit undesirable responses and change plans. However, scattered evidence implies that stopping behavior is entangled with simpler automatic stimulus-response mechanisms. Here we operationalize this idea by merging the latest conceptualization of saccadic countermanding with a neural network model of visuo-oculomotor behavior that integrates bottom-up and top-down drives. This model accounts for all fundamental qualitative and quantitative features of saccadic countermanding, including neuronal activity. Importantly, it does so by using the same architecture and parameters as basic visually guided behavior and automatic stimulus-driven interference. Using simulations and new data, we compare the temporal dynamics of saccade countermanding with that of saccadic inhibition (SI), a hallmark effect thought to reflect automatic competition within saccade planning areas. We demonstrate how SI accounts for a large proportion of the saccade countermanding process when using visual signals. We conclude that top-down inhibition acts later, piggy-backing on the quicker automatic inhibition. This conceptualization fully accounts for the known effects of signal features and response modalities traditionally used across the countermanding literature. Moreover, it casts different light on the concept of top-down inhibition, its timing and neural underpinning, as well as the interpretation of stop-signal reaction time (RT), the main behavioral measure in the countermanding literature. (PsycInfo Database Record (c) 2020 APA, all rights reserved).

Centroparietal activity mirrors the decision variable when tracking biased and time-varying sensory evidence

Decision-making is a fundamental human activity requiring explanation at the neurocognitive level. Current theoretical frameworks assume that, during sensory-based decision-making, the stimulus is sampled sequentially. The resulting evidence is accumulated over time as a decision variable until a threshold is reached and a response is initiated. Several neural signals, including the centroparietal positivity (CPP) measured from the human electroencephalogram (EEG), appear to display the accumulation-to-bound profile associated with the decision variable. Here, we evaluate the putative computational role of the CPP as a model-derived accumulation-to-bound signal, focussing on point-by-point correspondence between model predictions and data in order to go beyond simple summary measures like average slope. In two experiments, we explored the CPP under two manipulations (namely non-stationary evidence and probabilistic decision biases) that complement one another by targeting the shape and amplitude of accumulation respectively. We fit sequential sampling models to the behavioural data, and used the resulting parameters to simulate the decision variable, before directly comparing the simulated profile to the CPP waveform. In both experiments, model predictions deviated from our naïve expectations, yet showed similarities with the neurodynamic data, illustrating the importance of a formal modelling approach. The CPP appears to arise from brain processes that implement a decision variable (as formalised in sequential-sampling models) and may therefore inform our understanding of decision-making at both the representational and implementational levels of analysis, but at this point it is uncertain whether a single model can explain how the CPP varies across different kinds of task manipulation.

Visual search asymmetry depends on target-distractor feature similarity: Is the asymmetry simply a result of distractor rejection speed?

Previous studies have shown that in visual search, varying the target and distractor familiarity produces a search asymmetry: Detecting a novel target among familiar distractors is more efficient than detecting a familiar target among novel distractors. One explanation is that novel targets have enhanced salience and are detected preattentively. Conversely, familiar distractors may be easier to reject. The current study postulates that target-distractor feature similarity, in addition to target or distractor familiarity, is a key determinant of visual search efficiency. The results of two experiments reveal that visual search is more efficient when distractors are familiar regardless of target familiarity, but only when the target-distractor similarity is high. When similarity is low, the visual search asymmetry disappears and the search times become highly efficient, with search slopes not different from zero regardless of target or distractor familiarity. However, although distractor familiarity plays an important role in inducing the search asymmetry, comparisons of search efficiency in target-present and target-absent trials reveal that search asymmetries cannot be explained solely by the faster speed of rejecting familiar distractors, as proposed by previous studies. Rather, distractor familiarity influences processes outside of stimulus selection, such as search monitoring and termination decisions. Competition among bottom-up item salience effects and top-down shape recognition processes is proposed to account for these findings.

A divisive model of evidence accumulation explains uneven weighting of evidence over time

Divisive normalization has long been used to account for computations in various neural processes and behaviours. The model proposes that inputs into a neural system are divisively normalized by the system’s total activity. More recently, dynamical versions of divisive normalization have been shown to account for how neural activity evolves over time in value-based decision making. Despite its ubiquity, divisive normalization has not been studied in decisions that require evidence to be integrated over time. Such decisions are important when the information is not all available at once. A key feature of such decisions is how evidence is weighted over time, known as the integration kernel. Here, we provide a formal expression for the integration kernel in divisive normalization, and show that divisive normalization quantitatively accounts for 133 human participants’ perceptual decision making behaviour, performing as well as the state-of-the-art Drift Diffusion Model, the predominant model for perceptual evidence accumulation.

Divisive normalization is thought to be a ubiquitous computation in the brain, but has not been studied in decisions that require integrating evidence over time. Here, the authors show in humans that dynamic divisive normalization accounts for the uneven weighting of perceptual evidence over time.

Temporal cascade of frontal, motor and muscle processes underlying human action-stopping

Action-stopping is a canonical executive function thought to involve top-down control over the motor system. Here we aimed to validate this stopping system using high temporal resolution methods in humans. We show that, following the requirement to stop, there was an increase of right frontal beta (~13 to 30 Hz) at ~120 ms, likely a proxy of right inferior frontal gyrus; then, at 140 ms, there was a broad skeletomotor suppression, likely reflecting the impact of the subthalamic nucleus on basal ganglia output; then, at ~160 ms, suppression was detected in the muscle, and, finally, the behavioral time of stopping was ~220 ms. This temporal cascade supports a physiological model of action-stopping, and partitions it into subprocesses that are isolable to different nodes and are more precise than the behavioral latency of stopping. Variation in these subprocesses, including at the single-trial level, could better explain individual differences in impulse control.

Judging Relative Onsets and Offsets of Audiovisual Events

This study assesses the fidelity with which people can make temporal order judgments (TOJ) between auditory and visual onsets and offsets. Using an adaptive staircase task administered to a large sample of young adults, we find that the ability to judge temporal order varies widely among people, with notable difficulty created when auditory events closely follow visual events. Those findings are interpretable within the context of an independent channels model. Visual onsets and offsets can be difficult to localize in time when they occur within the temporal neighborhood of sound onsets or offsets.

Countermanding Perceptual Decision-Making

We investigated whether a task requiring concurrent perceptual decision-making and response control can be performed concurrently, whether evidence accumulation and response control are accomplished by the same neurons, and whether perceptual decision-making and countermanding can be unified computationally. Based on neural recordings in a prefrontal area of macaque monkeys, we present behavioral, neural, and computational results demonstrating that perceptual decision-making of varying difficulty can be countermanded efficiently, that single prefrontal neurons instantiate both evidence accumulation and response control, and that an interactive race between stochastic GO evidence accumulators for each alternative and a distinct STOP accumulator fits countermanding choice behavior and replicates neural trajectories. Thus, perceptual decision-making and response control, previously regarded as distinct mechanisms, are actually aspects of a common neuro-computational mechanism supporting flexible behavior.

Accumulators, Neurons, and Response Time

The marriage of cognitive neurophysiology and mathematical psychology to understand decision-making has been exceptionally productive. This interdisciplinary area is based on the proposition that particular neurons or circuits instantiate the accumulation of evidence specified by mathematical models of sequential sampling and stochastic accumulation. This linking proposition has earned widespread endorsement. Here, a brief survey of the history of the proposition precedes a review of multiple conundrums and paradoxes concerning the accuracy, precision, and transparency of that linking proposition. Correctly establishing how abstract models of decision-making are instantiated by particular neural circuits would represent a remarkable accomplishment in mapping mind to brain. Failing would reveal challenging limits for cognitive neuroscience. This is such a vigorous area of research because so much is at stake.

The role of premature evidence accumulation in making difficult perceptual decisions under temporal uncertainty

The computations and neural processes underpinning decision making have primarily been investigated using highly simplified tasks in which stimulus onsets cue observers to start accumulating choice-relevant information. Yet, in daily life we are rarely afforded the luxury of knowing precisely when choice-relevant information will appear. Here, we examined neural indices of decision formation while subjects discriminated subtle stimulus feature changes whose timing relative to stimulus onset ('foreperiod') was uncertain. Joint analysis of behavioural error patterns and neural decision signal dynamics indicated that subjects systematically began the accumulation process before any informative evidence was presented, and further, that accumulation onset timing varied systematically as a function of the foreperiod of the preceding trial. These results suggest that the brain can adjust to temporal uncertainty by strategically modulating accumulation onset timing according to statistical regularities in the temporal structure of the sensory environment with particular emphasis on recent experience.

Mutual benefits: Combining reinforcement learning with sequential sampling models

Reinforcement learning models of error-driven learning and sequential-sampling models of decision making have provided significant insight into the neural basis of a variety of cognitive processes. Until recently, model-based cognitive neuroscience research using both frameworks has evolved separately and independently. Recent efforts have illustrated the complementary nature of both modelling traditions and showed how they can be integrated into a unified theoretical framework, explaining trial-by-trial dependencies in choice behavior as well as response time distributions. Here, we review a theoretical background of integrating the two classes of models, and review recent empirical efforts towards this goal. We furthermore argue that the integration of both modelling traditions provides mutual benefits for both fields, and highlight promises of this approach for cognitive modelling and model-based cognitive neuroscience.

A method, framework, and tutorial for efficiently simulating models of decision-making

Evidence accumulation models (EAMs) have become the dominant models of rapid decision-making. Several variants of these models have been proposed, ranging from the simple linear ballistic accumulator (LBA) to the more complex leaky-competing accumulator (LCA), and further extensions that include time-varying rates of evidence accumulation or decision thresholds. Although applications of the simpler variants have been widespread, applications of the more complex models have been fewer, largely due to their intractable likelihood function and the computational cost of mass simulation. Here, I present a framework for efficiently fitting complex EAMs, which uses a new, efficient method of simulating these models. I find that the majority of simulation time is taken up by random number generation (RNG) from the normal distribution, needed for the stochastic noise of the differential equation. To reduce this inefficiency, I propose using the well-known concept within computer science of "look-up tables" (LUTs) as an approximation to the inverse cumulative density function (iCDF) method of RNG, which I call "LUT-iCDF". I show that when using an appropriately sized LUT, simulations using LUT-iCDF closely match those from the standard RNG method in R. My framework, which I provide a detailed tutorial on how to implement, includes C code for 12 different variants of EAMs using the LUT-iCDF method, and should make the implementation of complex EAMs easier and faster.