A sequential sampling account of semantic relatedness decisions

Semantic memory research often draws on decisions about the semantic relatedness of concepts. These decisions depend on cognitive processes of memory retrieval and choice formation. However, most previous research focused on memory retrieval but neglected the decision aspects. Here we propose the sequential sampling framework to account for choices and response times in semantic relatedness decisions. We focus on three popular sequential sampling models, the Race model, the Leaky Competing Accumulator model (LCA) and the Drift Diffusion Model (DDM). Using model simulations, we investigate if and how these models account for two empirical benchmarks: the relatedness effect, denoting faster "related" than "unrelated" decisions when judging the relatedness of word pairs; and an inverted-U shaped relationship between response time and the relatedness strength of word pairs. Our simulations show that the LCA and DDM, but not the Race model, can reproduce both effects. Furthermore, the LCA predicts a novel phenomenon: the inverted relatedness effect for weakly related word pairs. Reanalyzing a publicly available data set, we obtained credible evidence of such an inverted relatedness effect. These results provide strong support for sequential sampling models -- and in particular the LCA -- as a viable computational account of semantic relatedness decisions and suggest an important role for decision-related processes in (semantic) memory tasks.

Download Full-text

Aging and response times: a comparison of sequential sampling models

Measuring the Mind: Speed, control, and age ◽

10.1093/acprof:oso/9780198566427.003.0001 ◽

2005 ◽

pp. 3-32 ◽

Cited By ~ 9

Author(s):

Roger Ratcliff ◽

Anjali Thapar ◽

Philip L. Smith ◽

Gail McKoon

Keyword(s):

Response Times ◽

Sequential Sampling ◽

Sequential Sampling Models

Download Full-text

Response times in economics: Looking through the lens of sequential sampling models

Journal of Economic Psychology ◽

10.1016/j.joep.2018.09.008 ◽

2018 ◽

Vol 69 ◽

pp. 61-86 ◽

Cited By ~ 2

Author(s):

John A. Clithero

Keyword(s):

Response Times ◽

Sequential Sampling ◽

Sequential Sampling Models

Download Full-text

An LBA account of decisions in the multiple object tracking task

10.31234/osf.io/hkj2g ◽

2019 ◽

Author(s):

Reilly James Innes ◽

Caroline Kuhne

Keyword(s):

Decision Making ◽

Object Tracking ◽

Response Times ◽

Sequential Sampling ◽

Multiple Object Tracking ◽

Tracking Task ◽

Everyday Functioning ◽

Multiple Object ◽

Sequential Sampling Models ◽

Linear Ballistic Accumulator

Decision making is a vital aspect of our everyday functioning, from simple perceptual demands to more complex and meaningful decisions. The strategy adopted to make such decisions is often viewed as balancing elements of speed and caution, i.e. making fast or careful decisions. Using sequential sampling models to analyse decision making data can allow us to tease apart strategic differences, such as being more or less cautious, from processing differences, which would otherwise be indistinguishable in behavioural data. Our study used a multiple object tracking task where student participants and a highly skilled military group were compared on their ability to track several items at once. Using a mathematical model of decision making (the linear ballistic accumulator), we show the underpinnings of how two groups differ in performance. Results showed a large difference between the groups on accuracy, with the RAAF group outperforming students. An interaction effect was observed between groups and level of difficulty in response times, where RAAF response times slowed at a greater rate than the student group as difficulty increased. Model results indicated that the RAAF personnel were more cautious in their decisions than students, and had faster processing in some conditions. Our study shows the strength of sequential sampling models, as well as providing a first attempt at fitting a sequential sampling model to data from a multiple object tracking task.

Download Full-text

Unification of Countermanding and Perceptual Decision-making

10.1101/158170 ◽

2017 ◽

Author(s):

Paul G. Middlebrooks ◽

Bram B. Zandbelt ◽

Gordon D. Logan ◽

Thomas J. Palmeri ◽

Jeffrey D. Schall

Keyword(s):

Decision Making ◽

Response Inhibition ◽

Sequential Sampling ◽

Race Model ◽

Decision Time ◽

Perceptual Decision Making ◽

Perceptual Decision ◽

Evidence Accumulation ◽

Sequential Sampling Models ◽

Choice Tasks

Perceptual decision-making, studied using two-alternative forced-choice tasks, is explained by sequential sampling models of evidence accumulation, which correspond to the dynamics of neurons in sensorimotor structures of the brain1 2. Response inhibition, studied using stop-signal (countermanding) tasks, is explained by a race model of the initiation or canceling of a response, which correspond to the dynamics of neurons in sensorimotor structures3 4. Neither standard model accounts for performance of the other task. Sequential sampling models incorporate response initiation as an uninterrupted non-decision time parameter independent of task-related variables. The countermanding race model does not account for the choice process. Here we show with new behavioral, neural and computational results that perceptual decision making of varying difficulty can be countermanded with invariant efficiency, that single prefrontal neurons instantiate both evidence accumulation and response inhibition, and that an interactive race between two GO and one STOP stochastic accumulator fits countermanding choice behavior. Thus, perceptual decision-making and response control, previously regarded as distinct mechanisms, are actually aspects of more flexible behavior supported by a common neural and computational mechanism. The identification of this aspect of decision-making with response production clarifies the component processes of decision-making.

Download Full-text

Attention as a source of variability in decision-making: Accounting for overall-value effects with diffusion models

10.31234/osf.io/rewtq ◽

2020 ◽

Author(s):

Blair Shevlin ◽

Ian Krajbich

Keyword(s):

Decision Making ◽

Response Times ◽

Sequential Sampling ◽

Diffusion Models ◽

Tracking Data ◽

Sequential Sampling Models ◽

Relative Value ◽

Data Generating Process ◽

Multiplicative Relationship ◽

Value Effect

Research has demonstrated that value-based decisions depend not only on the relative value difference between options, but also on their overall value. In particular, response times tend to decrease as the overall value of the options increase. Standard sequential sampling models such as the diffusion model can account for this fact by assuming that decision thresholds or noise vary with overall value. Alternatively, attention-based models that incorporate eye-tracking data accommodate this overall-value effect directly as a consequence of the multiplicative relationship between attention and value magnitude. Using non-attentional diffusion models fit to data simulated with an attention-based model, we find that parameters related to decision thresholds or noise vary as a function of overall value, even though these were not features of the data generating process. We find additional evidence for misidentified parameters in a similar analysis using empirical data. Our results indicate that neglecting attentional effects can lead to mistaken conclusions about which decision parameters are sensitive to overall value.

Download Full-text

Sequential Sampling Models Without Random Between-Trial Variability: The Racing Diffusion Model of Speeded Decision

10.31234/osf.io/xuwbk ◽

2017 ◽

Cited By ~ 3

Author(s):

Gabriel Tillman

Keyword(s):

Response Time ◽

Diffusion Model ◽

Response Times ◽

Sequential Sampling ◽

Good Reason ◽

Accumulation Process ◽

Time Data ◽

Evidence Accumulation ◽

Sequential Sampling Models ◽

Sequential Sampling Model

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

Download Full-text

Evidence Accumulates for Individual Attributes during Value-Based Decisions

10.1101/2021.08.05.455296 ◽

2021 ◽

Author(s):

Douglas G. Lee ◽

Todd A. Hare

Keyword(s):

Computational Models ◽

Response Times ◽

Sequential Sampling ◽

Choice Data ◽

Attribute Importance ◽

Snack Foods ◽

Sequential Sampling Models ◽

Level Information ◽

Binary Choices ◽

Important Evidence

When choosing between different options, we tend to consider specific attribute qualities rather than deliberating over some general sense of the objects' overall values. The importance of each attribute together with its quality will determine our preference rankings over the available alternatives. Here, we show that the relative importance of the latent attributes within food rewards reliably differs when the items are evaluated in isolation compared to when binary choices are made between them. Specifically, we used standard regression and sequential sampling models to examine six datasets in which participants evaluated, and chose between, multi-attribute snack foods. We show that models that assume that attribute importance remains constant across evaluation and choice contexts fail to reproduce fundamental patterns in the choice data and provide quantitatively worse fits to the choice outcomes, response times, and confidence reports compared to models that allow for attribute importance to vary across preference elicitation methods. Our results provide important evidence that incorporating attribute-level information into computational models helps us to better understand the cognitive processes involved in value-based decision-making.

Download Full-text