Search

The cognitive architecture ACT-R is at the same time a psychological theory and a modeling framework for constructing cognitive models that adhere to the principles of the theory. ACT-R can be used in combination with fMRI data in two different ways: (1) fMRI data can be used to evaluate and constrain models in ACT-R by means of predefined Region-of-Interest (ROI) analysis, and (2) predictions from ACT-R models can be used to locate neural correlates of model processes and representations by means of model-based fMRI analysis. In this paper we provide a step-by-step tutorial on both approaches. Note that this tutorial neither teaches the ACT-R theory in any detail, nor fMRI analysis, but explains how ACT-R can be used in combination with fMRI data. To this end, we provide all data and computer code necessary to run the ACT-R model, carry out the analyses, and recreate the figures in the paper. As an example dataset we use a relatively simple algebra task. In the first section, we develop an ACT-R model of this task and fit it to behavioral data. In the second section, we apply a predefined ROI-analysis to evaluate the model using fMRI data. In the third section, we use model-based fMRI analysis to locate the following processes in the brain: retrieval of mathematical facts from memory, working memory updates, motor responses, and visually encoding the problems. After working through this tutorial, the reader will have learned what can be achieved with the two different analysis methods and how they are conducted; the example code can then be adapted to a new dataset. [ABSTRACT FROM AUTHOR]

Reinforcement learning (RL) models have been widely used to analyze the choice behavior of humans and other animals in a broad range of fields, including psychology and neuroscience. Linear regression-based models that explicitly represent how reward and choice history influences future choices have also been used to model choice behavior. While both approaches have been used independently, the relation between the two models has not been explicitly described. The aim of the present study is to describe this relation and investigate how the parameters in the RL model mediate the effects of reward and choice history on future choices. To achieve these aims, we performed analytical calculations and numerical simulations. First, we describe a special case in which the RL and regression models can provide equivalent predictions of future choices. The general properties of the RL model are discussed as a departure from this special case. We clarify the role of the RL-model parameters, specifically, the learning rate, inverse temperature, and outcome value (also referred to as the reward value, reward sensitivity, or motivational value), in the formation of history dependence. [ABSTRACT FROM AUTHOR]

Abstract: The aim is to investigate the consistency or variability of catchment response over time and space and evaluate the predictive error caused by the impacts of climate variability on streamflow. For this purpose, both data- and top-down model-based analyses of the dynamic relation between rainfall and runoff for selected sub-catchments have been undertaken. Data analysis techniques (e.g. trend analysis, deconvolution and baseflow filtering) were used to assess the temporal and spatial variation in the hydrologic response characteristics for each site. The lumped conceptual rainfall–runoff model IHACRES CMD (Catchment Moisture Deficit) version is applied to the sub-catchments to assess the adequacy of the model response in representing the impact of weather patterns on streamflow. Several performance criteria have been used to evaluate the performance of the model in each calibration period using a multi-criteria approach. The IHACRES-3S (3 Storage) model is applied to assess low flow behaviour and capture the timing in the switch between baseflow and no flow periods. Rainfall–runoff model performance characteristics of each sub-catchment are quite related to their incident rainfall regime. Sub-catchments which are located in a lower rainfall regime show poor to average model performance. The reduction in performance in R 2 is due to the poor fitting to the peaks for both large and small streamflow events, with the model underestimating the highest flow peaks, and overestimating smaller peaks. Further work will be needed to assess observed data reliability and improve model performance in order to separate the impacts of climate variations and land use change on hydrological response. An appropriate model structure having a variable partitioning between quick and slow flow components is under consideration and techniques are being used to identify problematic periods and events with high error in the observational data. [Copyright &y& Elsevier]