Qualitative speed-accuracy tradeoff effects that cannot be explained by the diffusion model under the selective influence assumption

The neural network RTNet exhibits the signatures of human perceptual decision-making

Convolutional neural networks show promise as models of biological vision. However, their decision behaviour, including the facts that they are deterministic and use equal numbers of computations for easy and difficult stimuli, differs markedly from human decision-making, thus limiting their applicability as models of human perceptual behaviour. Here we develop a new neural network, RTNet, that generates stochastic decisions and human-like response time (RT) distributions. We further performed comprehensive tests that showed RTNet reproduces all foundational features of human accuracy, RT and confidence and does so better than all current alternatives. To test RTNet's ability to predict human behaviour on novel images, we collected accuracy, RT and confidence data from 60 human participants performing a digit discrimination task. We found that the accuracy, RT and confidence produced by RTNet for individual novel images correlated with the same quantities produced by human participants. Critically, human participants who were more similar to the average human performance were also found to be closer to RTNet's predictions, suggesting that RTNet successfully captured average human behaviour. Overall, RTNet is a promising model of human RTs that exhibits the critical signatures of perceptual decision-making.

A low-dimensional approximation of optimal confidence

Human decision making is accompanied by a sense of confidence. According to Bayesian decision theory, confidence reflects the learned probability of making a correct response, given available data (e.g., accumulated stimulus evidence and response time). Although optimal, independently learning these probabilities for all possible data combinations is computationally intractable. Here, we describe a novel model of confidence implementing a low-dimensional approximation of this optimal yet intractable solution. This model allows efficient estimation of confidence, while at the same time accounting for idiosyncrasies, different kinds of biases and deviation from the optimal probability correct. Our model dissociates confidence biases resulting from the estimate of the reliability of evidence by individuals (captured by parameter α), from confidence biases resulting from general stimulus independent under and overconfidence (captured by parameter β). We provide empirical evidence that this model accurately fits both choice data (accuracy, response time) and trial-by-trial confidence ratings simultaneously. Finally, we test and empirically validate two novel predictions of the model, namely that 1) changes in confidence can be independent of performance and 2) selectively manipulating each parameter of our model leads to distinct patterns of confidence judgments. As a tractable and flexible account of the computation of confidence, our model offers a clear framework to interpret and further resolve different forms of confidence biases.

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.

Computational analysis of speed-accuracy tradeoff

Speed-accuracy tradeoff (SAT) in the decision making of humans and animals is a well-documented phenomenon, but its underlying neuronal mechanism remains unclear. Modeling approaches have conceptualized SAT through the threshold hypothesis as adjustments to the decision threshold. However, the leading neurophysiological view is the gain modulation hypothesis. This hypothesis postulates that the SAT mechanism is implemented through changes in the dynamics of the choice circuit, which increase the baseline firing rate and the speed of neuronal integration. In this paper, I investigated alternative computational mechanisms of SAT and showed that the threshold hypothesis was qualitatively consistent with the behavioral data, but the gain modulation hypothesis was not. In order to reconcile the threshold hypothesis with the neurophysiological evidence, I considered the interference of alpha oscillations with the decision process and showed that alpha oscillations could increase the discriminatory power of the decision system, although they slowed down the decision process. This suggests that the magnitude of alpha waves suppression during the event related desynchronization (ERD) of alpha oscillations depends on a SAT condition and the amplitude of alpha oscillations is lower in the speed condition. I also showed that the lower amplitude of alpha oscillations resulted in an increase in the baseline firing rate and the speed of neuronal intergration. Thus, the interference of the event related desynchronization of alpha oscillations with a SAT condition explains why an increase in the baseline firing rate and the speed of neuronal integration accompany the speed condition.

Motivation by reward jointly improves speed and accuracy, whereas task-relevance and meaningful images do not

Visual selection is characterized by a trade-off between speed and accuracy. Speed or accuracy of the selection process can be affected by higher level factors—for example, expecting a reward, obtaining task-relevant information, or seeing an intrinsically relevant target. Recently, motivation by reward has been shown to simultaneously increase speed and accuracy, thus going beyond the speed–accuracy-trade-off. Here, we compared the motivating abilities of monetary reward, task-relevance, and image content to simultaneously increase speed and accuracy. We used a saccadic distraction task that required suppressing a distractor and selecting a target. Across different blocks successful target selection was followed either by (i) a monetary reward, (ii) obtaining task-relevant information, or (iii) seeing the face of a famous person. Each block additionally contained the same number of irrelevant trials lacking these consequences, and participants were informed about the upcoming trial type. We found that postsaccadic vision of a face affected neither speed nor accuracy, suggesting that image content does not affect visual selection via motivational mechanisms. Task relevance increased speed but decreased selection accuracy, an observation compatible with a classical speed–accuracy trade-off. Motivation by reward, however, simultaneously increased response speed and accuracy. Saccades in all conditions deviated away from the distractor, suggesting that the distractor was suppressed, and this deviation was strongest in the reward block. Drift-diffusion modelling revealed that task-relevance affected behavior by affecting decision thresholds, whereas motivation by reward additionally increased the rate of information uptake. The present findings thus show that the three consequences differ in their motivational abilities.

An adaptive paradigm for detecting the individual duration of the preparatory period in the choice reaction time task

According to the sequential stage model, the selection and the execution of a motor response are two distinct independent processes. Here, we propose a new adaptive paradigm for identifying the individual duration of the response preparatory period based on the motor reaction time (RT) data. The results are compared using the paradigm with constant values of the preparatory period. Two groups of participants performed on either an easy (Group 1) or a hard (Group 2) response selection task with two types of stimuli based on the preparatory period parameters: (1) stimuli with a constant preparatory period duration of 0 or 1200 ms and (2) stimuli with adaptive preparatory period durations. Our analysis showed an increase in the duration of the response selection process as a function of increasing task complexity when using both paradigms with constant and adaptive values of the preparatory period duration. We conclude that the adaptive paradigm proposed in the current paper has several important advantages over the constant paradigm in terms of measuring the response accuracy while being equally efficiently in capturing other critical response parameters.

Dynamic influences on static measures of metacognition

Humans differ in their capability to judge choice accuracy via confidence judgments. Popular signal detection theoretic measures of metacognition, such as M-ratio, do not consider the dynamics of decision making. This can be problematic if response caution is shifted to alter the tradeoff between speed and accuracy. Such shifts could induce unaccounted-for sources of variation in the assessment of metacognition. Instead, evidence accumulation frameworks consider decision making, including the computation of confidence, as a dynamic process unfolding over time. Using simulations, we show a relation between response caution and M-ratio. We then show the same pattern in human participants explicitly instructed to focus on speed or accuracy. Finally, this association between M-ratio and response caution is also present across four datasets without any reference towards speed. In contrast, when data are analyzed with a dynamic measure of metacognition, v-ratio, there is no effect of speed-accuracy tradeoff.

Patterns of Movement Performance and Consistency From Childhood to Old Age

It is widely accepted that the general process of aging can be reflected by changes in motor function. Typically, optimal performance of a given motor task is observed for healthy young adults with declines being observed for individuals at either end of the lifespan. This study was designed to examine differences in the average and variability (i.e., intraindividual variability) of chewing, simple reaction time, postural control, and walking responses. For this study, 15 healthy children, 15 young adults, and 15 older adults participated. Our results indicated the movement performance for the reaction time and postural sway followed a U shape with young adults having faster reaction times and decreased postural sway compared to the children and older adults. However, this pattern was not preserved across all motor tasks with no age differences emerging for (normalized) gait speed, while chewing rates followed a U-shaped curve with older adults and children chewing at faster rates. Taken together, these findings would indicate that the descriptive changes in motor function with aging are heavily influenced by the nature of the task being performed and are unlikely to follow a singular pattern.

Understanding neural signals of post-decisional performance monitoring: An integrative review

Performance monitoring is a key cognitive function, allowing to detect mistakes and adapt future behavior. Post-decisional neural signals have been identified that are sensitive to decision accuracy, decision confidence and subsequent adaptation. Here, we review recent work that supports an understanding of late error/confidence signals in terms of the computational process of post-decisional evidence accumulation. We argue that the error positivity, a positive-going centro-parietal potential measured through scalp electrophysiology, reflects the post-decisional evidence accumulation process itself, which follows a boundary crossing event corresponding to initial decision commitment. This proposal provides a powerful explanation for both the morphological characteristics of the signal and its relation to various expressions of performance monitoring. Moreover, it suggests that the error positivity -a signal with thus far unique properties in cognitive neuroscience - can be leveraged to furnish key new insights into the inputs to, adaptation, and consequences of the post-decisional accumulation process.

Qualitative speed-accuracy tradeoff effects can be explained by a diffusion/fast-guess mixture model

Rafiei and Rahnev (2021) presented an analysis of an experiment in which they manipulated speed-accuracy stress and stimulus contrast in an orientation discrimination task. They argued that the standard diffusion model could not account for the patterns of data their experiment produced. However, their experiment encouraged and produced fast guesses in the higher speed-stress conditions. These fast guesses are responses with chance accuracy and response times (RTs) less than 300 ms. We developed a simple mixture model in which fast guesses were represented by a simple normal distribution with fixed mean and standard deviation and other responses by the standard diffusion process. The model fit the whole pattern of accuracy and RTs as a function of speed/accuracy stress and stimulus contrast, including the sometimes bimodal shapes of RT distributions. In the model, speed-accuracy stress affected some model parameters while stimulus contrast affected a different one showing selective influence. Rafiei and Rahnev's failure to fit the diffusion model was the result of driving subjects to fast guess in their experiment.