Analysis of the human connectome data supports the notion of a "Common Model of Cognition" for human and human-like intelligence across domains.

The Common Model of Cognition (CMC) is a recently proposed, consensus architecture intended to capture decades of progress in cognitive science on modeling human and human-like intelligence. Because of the broad agreement around it and preliminary mappings of its components to specific brain areas, we hypothesized that the CMC could be a candidate model of the large-scale functional architecture of the human brain. To test this hypothesis, we analyzed functional MRI data from 200 participants and seven different tasks that cover a broad range of cognitive domains. The CMC components were identified with functionally homologous brain regions through canonical fMRI analysis, and their communication pathways were translated into predicted patterns of effective connectivity between regions. The resulting dynamic linear model was implemented and fitted using Dynamic Causal Modeling, and compared against six alternative brain architectures that had been previously proposed in the field of neuroscience (three hierarchical architectures and three hub-and-spoke architectures) using a Bayesian approach. The results show that, in all cases, the CMC vastly outperforms all other architectures, both within each domain and across all tasks. These findings suggest that a common set of architectural principles that could be used for artificial intelligence also underpins human brain function across multiple cognitive domains.

The bilingual language network: Differential involvement of anterior cingulate, basal ganglia and prefrontal cortex in preparation, monitoring, and execution.

Research on the neural bases of bilingual language control has largely overlooked the role of preparatory processes, which are central to cognitive control. Additionally, little is known about how the processes involved in global language selection may differ from those involved in the selection of words and morpho-syntactic rules for manipulating them. These processes were examined separately in an fMRI experiment, with an emphasis on understanding how and when general cognitive control regions become activated. Results of region-of-interest analyses on 23 early Spanish-English bilinguals showed that the anterior cingulate cortex (ACC) was primarily engaged during the language preparation phase of the task, whereas the left prefrontal (DLPFC) and pre-supplementary motor areas showed increasing activation from preparation to execution. Activation in the basal ganglia (BG), left middle temporal lobe, and right precentral cortical regions did not significantly differ throughout the task. These results suggest that three core cognitive control regions, the ACC, DLPFC, and BG, which have been previously implicated in bilingual language control, engage in distinct neurocognitive processes. Specifically, the results are consistent with the view that the BG "keep track" of the target language in use throughout various levels of language selection, that the ACC is particularly important for top-down target language preparation, and that the left prefrontal cortex is increasingly involved in selection processes from preparation through task execution.

Inhibitory synapses between striatal projection neurons support efficient enhancement of cortical signals: a computational model.

The function of lateral inhibitory synapses between striatal projection neurons is currently poorly understood. This paper puts forward a model suggesting that inhibitory collaterals can be used to enhance the incoming cortical signals. In particular, we propose that lateral inhibition between projection neurons performs a signal-enhancing process that resembles the image processing technique of "unsharp masking", where a blurred copy is used to enhance and sharpen an input image. The paper also presents the results of computer simulations deomsntrating that the proposed mechanisms is compatible with known properties of striatal projection neurons, and outperforms alternative models of lateral inhibition. Finally, this paper illustrates the advantages of the proposed model and discusses the relevance of these conclusions for existing computational models of the basal ganglia and their role in cognition.

Allocating Mental Effort in Cognitive Tasks: A Model of Motivation in the ACT-R Cognitive Architecture.

Motivation is the driving force that influences people's behaviors and interacts with many cognitive functions. Computationally, motivation is represented as a cost-benefit analysis that weighs efforts and rewards in order to choose the optimal actions. Shenhav and colleagues proposed an elegant theory, the Expected Value of Control (EVC), which describes the relationship between cognitive efforts, costs, and rewards. In this paper, we propose a more fine-grained and detailed motivation framework that incorporates the principles of EVC into the ACT-R cognitive architecture. Specifically, motivation is represented as a specific slot in the Goal buffer with a corresponding scalar value, M, that is translated into the reward value Rt that is delivered when the goal is reached. This implementation is tested in two models. The first model is a high-level model that reproduces the EVC predictions with abstract actions. The second model is an augmented version of an existing ACT-R model of the Simon task. The motivation mechanism is shown to permit optimal effort allocation and reproduce known phenomena. Finally, the broader implications of our mechanism are discussed.

Reliance on Episodic vs. Procedural Systems in Decision-Making Depends on Individual Differences in Their Relative Neural Efficiency.

Experiential decision-making can be explained as a result of either memory-based or reinforcement-based processes. Here, for the first time, we show that individual preferences between a memory-based and a reinforcement-based strategy, even when the two are functionally equivalent in terms of expected payoff, are adaptively shaped by individual differences in resting-state brain connectivity between the corresponding brain regions. Using computational cognitive models to identify which mechanism was most likely used by each participant, we found that individuals with comparatively stronger connectivity between memory regions prefer a memory-based strategy, while individuals with comparatively stronger connectivity between sensorimotor and habit-formation regions preferentially rely on a reinforcement-based strategy. These results suggest that human decision-making is adaptive and sensitive to the neural costs associated with different strategies.

One Size Does Not Fit All: Idiographic Computational Models Reveal Individual Differences in Learning and Meta-Learning Strategies.

Complex skill learning depends on the joint contribution of multiple interacting systems: working memory (WM), declarative long-term memory (LTM) and reinforcement learning (RL). The present study aims to understand individual differences in the relative contributions of these systems during learning. We built four idiographic, ACT-R models of performance on the stimulus-response learning, Reinforcement Learning Working Memory task. The task consisted of short 3-image, and long 6-image, feedback-based learning blocks. A no-feedback test phase was administered after learning, with an interfering task inserted between learning and test. Our four models included two single-mechanism RL and LTM models, and two integrated RL-LTM models: (a) RL-based meta-learning, which selects RL or LTM to learn based on recent success, and (b) a parameterized RL-LTM selection model at fixed proportions independent of learning success. Each model was the best fit for some proportion of our learners (LTM: 68.7%, RL: 4.8%, Meta-RL: 13.25%, bias-RL:13.25% of participants), suggesting fundamental differences in the way individuals deploy basic learning mechanisms, even for a simple stimulus-response task. Finally, long-term declarative memory seems to be the preferred learning strategy for this task regardless of block length (3- vs 6-image blocks), as determined by the large number of subjects whose learning characteristics were best captured by the LTM only model, and a preference for LTM over RL in both of our integrated-models, owing to the strength of our idiographic approach.

Rapid instructed task learning: a new window into the human brain's unique capacity for flexible cognitive control.

The human ability to flexibly adapt to novel circumstances is extraordinary. Perhaps the most illustrative, yet underappreciated, form of this cognitive flexibility is rapid instructed task learning (RITL)--the ability to rapidly reconfigure our minds to perform new tasks from instructions. This ability is important for everyday life (e.g., learning to use new technologies) and is used to instruct participants in nearly every study of human cognition. We review the development of RITL as a circumscribed domain of cognitive neuroscience investigation, culminating in recent demonstrations that RITL is implemented via brain circuits centered on lateral prefrontal cortex. We then build on this and the recent discovery of compositional representations within lateral prefrontal cortex to develop an integrative theory of cognitive flexibility and cognitive control that identifies mechanisms that may enable RITL within the human brain. The insights gained from this new theoretical account have important implications for further developments and applications of RITL research.

Distinct contributions of the caudate nucleus, rostral prefrontal cortex, and parietal cortex to the execution of instructed tasks.

When we behave according to rules and instructions, our brains interpret abstract representations of what to do and transform them into actual behavior. In order to investigate the neural mechanisms behind this process, we devised an fMRI experiment that explicitly isolated rule interpretation from rule encoding and execution. Our results showed that a specific network of regions (including the left rostral prefrontal cortex, the caudate nucleus, and the bilateral posterior parietal cortices) is responsible for translating rules into executable form. An analysis of activation patterns across conditions revealed that the posterior parietal cortices represent a mental template for the task to perform, that the inferior parietal gyrus and the caudate nucleus are responsible for instantiating the template in the proper context, and that the left rostral prefrontal cortex integrates information across complex relationships.

Confidence-driven weighted retraining for predicting safety-critical failures in autonomous driving systems.

Safe handling of hazardous driving situations is a task of high practical relevance for building reliable and trustworthy cyber-physical systems such as autonomous driving systems. This task necessitates an accurate prediction system of the vehicle's confidence to prevent potentially harmful system failures on the occurrence of unpredictable conditions that make it less safe to drive. In this paper, we discuss the challenges of adapting a misbehavior predictor with knowledge mined during the execution of the main system. Then, we present a framework for the continual learning of misbehavior predictors, which records in-field behavioral data to determine what data are appropriate for adaptation. Our framework guides adaptive retraining using a novel combination of in-field confidence metric selection and reconstruction error-based weighing. We evaluate our framework to improve a misbehavior predictor from the literature on the Udacity simulator for self-driving cars. Our results show that our framework can reduce the false positive rate by a large margin and can adapt to nominal behavior drifts while maintaining the original capability to predict failures up to several seconds in advance.

The Structured Mind at Rest: Low-Frequency Oscillations Reflect Interactive Dynamics Between Spontaneous Brain Activity and a Common Architecture for Task Control.

The Common Model of Cognition (CMC) has been proposed as a high level framework through which functional neuroimaging data can be predicted and interpreted. Previous work has found the CMC is capable of predicting brain activity across a variety of tasks, but it has not been tested on resting state data. This paper adapts a previously used method for comparing theoretical models of brain structure, Dynamic Causal Modeling, for the task-free environment of resting state, and compares the CMC against six alternate architectural frameworks while also separately modeling spontaneous low-frequency oscillations. For a large sample of subjects from the Human Connectome Project, the CMC provides the best account of resting state brain activity, suggesting the presence of a general purpose structure of connections in the brain that drives activity when at rest and when performing directed task behavior. At the same time, spontaneous brain activity was found to be present and significant across all frequencies and in all regions. Together, these results suggest that, at rest, spontaneous low-frequency oscillations interact with the general cognitive architecture for task-based activity. The possible functional implications of these findings are discussed.

Inferring a Cognitive Architecture from Multitask Neuroimaging Data: A Data-Driven Test of the Common Model of Cognition Using Granger Causality.

Cognitive architectures (i.e., theorized blueprints on the structure of the mind) can be used to make predictions about the effect of multiregion brain activity on the systems level. Recent work has connected one high-level cognitive architecture, known as the "Common Model of Cognition," to task-based functional MRI data with great success. That approach, however, was limited in that it was intrinsically top-down, and could thus only be compared with alternate architectures that the experimenter could contrive. In this paper, we propose a bottom-up method to infer a cognitive architecture directly from brain imaging data itself, overcoming this limitation. Specifically, Granger causality modeling was applied to the same task-based fMRI data to infer a network of causal connections between brain regions based on their functional connectivity. The resulting network shares many connections with those proposed by the Common Model of Cognition but also suggests important additions likely related to the role of episodic memory. This combined top-down and bottom-up modeling approach can be used to help formalize the computational instantiation of cognitive architectures and further refine a comprehensive theory of cognition.

Increased Basal Ganglia Modulatory Effective Connectivity Observed in Resting-State fMRI in Individuals With Parkinson's Disease.

Alterations to interactions between networked brain regions underlie cognitive impairment in many neurodegenerative diseases, providing an important physiological link between brain structure and cognitive function. Previous attempts to characterize the effects of Parkinson's disease (PD) on network functioning using resting-state functional magnetic resonance imaging (rs-fMRI), however, have yielded inconsistent and contradictory results. Potential problems with prior work arise in the specifics of how the area targeted by the diseases (the basal ganglia) interacts with other brain regions. Specifically, current computational models point to the fact that the basal ganglia contributions should be captured with modulatory (i.e., second-order) rather than direct (i.e., first-order) functional connectivity measures. Following this hypothesis, a principled but manageable large-scale brain architecture, the Common Model of Cognition, was used to identify differences in basal ganglia connectivity in PD by analyzing resting-state fMRI data from 111 participants (70 patients with PD; 41 healthy controls) using Dynamic Causal Modeling (DCM). Specifically, the functional connectivity of the basal ganglia was modeled as two second-level, modulatory connections that control projections from sensory cortices to the prefrontal cortex, and from the hippocampus and medial temporal lobe to the prefrontal cortex. We then examined group differences between patients with PD and healthy controls in estimated modulatory effective connectivity in these connections. The Modulatory variant of the Common Model of Cognition outperformed the Direct model across all subjects. It was also found that these second-level modulatory connections had higher estimates of effective connectivity in the PD group compared to the control group, and that differences in effective connectivity were observed for all direct connections between the PD and control groups.We make the case that accounting for modulatory effective connectivity better captures the effects of PD on network functioning and influences the interpretation of the directionality of the between-group results. Limitations include that the PD group was scanned on dopaminergic medication, results were derived from a reasonable but small number of individuals and the ratio of PD to healthy control participants was relatively unbalanced. Future research will examine if the observed effect holds for individuals with PD scanned off their typical dopaminergic medications.

Qualitative Invariant Effects Arise from Neural Constraints: Common Architecture and Sources of Individual Differences.

The distinction between qualitative and quantitative effects can be meaningfully related to the distinction between "architecture" and "model" in computational cognitive science. In turn, this distinction can be related to differences between the invariant systems-level organization of the human brain and individual differences in structural and functional activity. Taken together, this presents an iterative new way to answer Newell's "20 Questions" problem and to systematize psychological effects as belonging to either architecture or the individual variations within it. Although some limits to this approach exist (for example, Individual differences in strategy might affect whether an effect is qualitative or quantitative), the approach might also shed light on how to account for special populations (neurological or psychological patients) within the same framework.

Recovering Reliable Idiographic Biological Parameters from Noisy Behavioral Data: the Case of Basal Ganglia Indices in the Probabilistic Selection Task.

Behavioral data, despite being a common index of cognitive activity, is under scrutiny for having poor reliability as a result of noise or lacking replications of reliable effects. Here, we argue that cognitive modeling can be used to enhance the test-retest reliability of the behavioral measures by recovering individual-level parameters from behavioral data. We tested this empirically with the Probabilistic Stimulus Selection (PSS) task, which is used to measure a participant's sensitivity to positive or negative reinforcement. An analysis of 400,000 simulations from an Adaptive Control of Thought-Rational (ACT-R) model of this task showed that the poor reliability of the task is due to the instability of the end-estimates: because of the way the task works, the same participants might sometimes end up having apparently opposite scores. To recover the underlying interpretable parameters and enhance reliability, we used a Bayesian Maximum A Posteriori (MAP) procedure. We were able to obtain reliable parameters across sessions (intraclass correlation coefficient ≈ 0.5). A follow-up study on a modified version of the task also found the same pattern of results, with very poor test-retest reliability in behavior but moderate reliability in recovered parameters (intraclass correlation coefficient ≈ 0.4). Collectively, these results imply that this approach can further be used to provide superior measures in terms of reliability, and bring greater insights into individual differences.

Core Cognitive Mechanisms Underlying Syntactic Priming: A Comparison of Three Alternative Models.

Syntactic priming (SP) is the effect by which, in a dialogue, the current speaker tends to re-use the syntactic constructs of the previous speakers. SP has been used as a window into the nature of syntactic representations within and across languages. Because of its importance, it is crucial to understand the mechanisms behind it. Currently, two competing theories exist. According to the transient activation account, SP is driven by the re-activation of declarative memory structures that encode structures. According to the error-based implicit learning account, SP is driven by prediction errors while processing sentences. By integrating both transient activation and associative learning, Reitter et al.'s hybrid model 2011 assumes that SP is achieved by both mechanisms, and predicts a priming enhancement for rare or unusual constructions. Finally, a recently proposed account, the reinforcement learning account, claims that SP driven by the successful application of procedural knowledge will be reversed when the prime sentence includes grammatical errors. These theories make different assumptions about the representation of syntactic rules (declarative vs. procedural) and the nature of the mechanism that drives priming (frequency and repetition, attention, and feedback signals, respectively). To distinguish between these theories, they were all implemented as computational models in the ACT-R cognitive architecture, and their specific predictions were examined through grid-search computer simulations. Two experiments were then carried out to empirically test the central prediction of each theory as well as the individual fits of each participant's responses to different parameterizations of each model. The first experiment produced results that were best explained by the associative account, but could also be accounted for by a modified reinforcement model with a different parsing algorithm. The second experiment, whose stimuli were designed to avoid the parsing ambiguity of the first, produced somewhat weaker effects. Its results, however, were also best predicted by the model implementing the associative account. We conclude that the data overall points to SP being due to prediction violations that direct attentional resources, in turn suggesting a declarative rather than a RL based procedural representation of syntactic rules.

Individual Differences in Reward-Based Learning Predict Fluid Reasoning Abilities.

The ability to reason and problem-solve in novel situations, as measured by the Raven's Advanced Progressive Matrices (RAPM), is highly predictive of both cognitive task performance and real-world outcomes. Here we provide evidence that RAPM performance depends on the ability to reallocate attention in response to self-generated feedback about progress. We propose that such an ability is underpinned by the basal ganglia nuclei, which are critically tied to both reward processing and cognitive control. This hypothesis was implemented in a neurocomputational model of the RAPM task, which was used to derive novel predictions at the behavioral and neural levels. These predictions were then verified in one neuroimaging and two behavioral experiments. Furthermore, an effective connectivity analysis of the neuroimaging data confirmed a role for the basal ganglia in modulating attention. Taken together, these results suggest that individual differences in a neural circuit related to reward processing underpin human fluid reasoning abilities.

Reflections of idiographic long-term memory characteristics in resting-state neuroimaging data.

Translational applications of cognitive science depend on having predictive models at the individual, or idiographic, level. However, idiographic model parameters, such as working memory capacity, often need to be estimated from specific tasks, making them dependent on task-specific assumptions. Here, we explore the possibility that idiographic parameters reflect an individual's biology and can be identified from task-free neuroimaging measures. To test this hypothesis, we correlated a reliable behavioral trait, the individual rate of forgetting in long-term memory, with a readily available task-free neuroimaging measure, the resting-state EEG spectrum. Using an established, adaptive fact-learning procedure, the rate of forgetting for verbal and visual materials was measured in a sample of 50 undergraduates from whom we also collected eyes-closed resting-state EEG data. Statistical analyses revealed that the individual rates of forgetting were significantly correlated across verbal and visual materials. Importantly, both rates correlated with resting-state power levels in the low (13-15 Hz) and upper (15-17 Hz) portion of the beta frequency bands. These correlations were particularly strong for visuospatial materials, were distributed over multiple fronto-parietal locations, and remained significant even after a correction for multiple comparisons (False Discovery Rate) and after robust correlation methods were applied. These results suggest that computational models could be individually tailored for prediction using idiographic parameter values derived from inexpensive, task-free imaging recordings.

When Fear Shrinks the Brain: A Computational Model of the Effects of Posttraumatic Stress on Hippocampal Volume.

Post-traumatic stress disorder (PTSD) is a psychiatric disorder often characterized by the unwanted re-experiencing of a traumatic event through nightmares, flashbacks, and/or intrusive memories. This paper presents a neurocomputational model using the ACT-R cognitive architecture that simulates intrusive memory retrieval following a potentially traumatic event (PTE) and predicts hippocampal volume changes observed in PTSD. Memory intrusions were captured in the ACT-R rational analysis framework by weighting the posterior probability of re-encoding traumatic events into memory with an emotional intensity term I to capture the degree to which an event was perceived as dangerous or traumatic. It is hypothesized that (1) increasing the intensity I of a PTE will increase the odds of memory intrusions, and (2) increased frequency of intrusions will result in a concurrent decrease in hippocampal size. A series of simulations were run and it was found that I had a significant effect on the probability of experiencing traumatic memory intrusions following a PTE. The model also found that I was a significant predictor of hippocampal volume reduction, where the mean and range of simulated volume loss match results of existing meta-analyses. The authors believe that this is the first model to both describe traumatic memory retrieval and provide a mechanistic account of changes in hippocampal volume, capturing one plausible link between PTSD and hippocampal volume.

The Role of Basal Ganglia Reinforcement Learning in Lexical Ambiguity Resolution.

The current study aimed to elucidate the contributions of the subcortical basal ganglia to human language by adopting the view that these structures engage in a basic neurocomputation that may account for its involvement across a wide range of linguistic phenomena. Specifically, we tested the hypothesis that basal ganglia reinforcement learning (RL) mechanisms may account for variability in semantic selection processes necessary for ambiguity resolution. To test this, we used a biased homograph lexical ambiguity priming task that allowed us to measure automatic processes for resolving ambiguity toward high-frequency word meanings. Individual differences in task performance were then related to indices of basal ganglia RL, which were used to group subjects into three learning styles: (a) Choosers who learn by seeking high reward probability stimuli; (b) Avoiders, who learn by avoiding low reward probability stimuli; and (c) Balanced participants, whose learning reflects equal contributions of choose and avoid processes. The results suggest that balanced individuals had significantly lower access to subordinate, or low-frequency, homograph word meanings. Choosers and Avoiders, on the other hand, had higher access to the subordinate word meaning even after a long delay between prime and target. Experimental findings were then tested using an ACT-R computational model of RL that learns from both positive and negative feedback. Results from the computational model simulations confirm and extend the pattern of behavioral findings, providing an RL account of individual differences in lexical ambiguity resolution.

Monitoring of attentional oscillations through Spectral Similarity Analysis predicts reading comprehension.

Deviations of attention from the task at hand are often associated with worse reading performance (Schooler, Reichle, & Halpern, 2004). Ironically, current methods for detecting these shifts of attention typically generate task interruptions and further disrupt performance. In the current study, we developed a method to (1) track shifts of attention away from the reading task by examining the similarity between 5 min of eyes-closed-resting-state EEG and 5 min reading EEG; and (2) investigate, during reading, how the ratio between attention shifts and focused reading relates to readers' comprehension. We performed a Spectral Similarity Analysis (SSA) that examined the spectral similarity between EEG recorded during reading and at rest on a moment-by-moment basis. We then recursively applied the algorithm to the resting-state data itself to obtain an individual baseline of the stability of brain activation recorded during rest. We defined any moment in which SSA during reading was greater than the mean correlation between resting-state EEG and itself as an "attentional shift." The results showed that the proportion of such attentional shifts recorded over the left visual region (O1) significantly predicted reading comprehension, with higher ratios (indicative of more frequent attentional shifts) relating to worse comprehension scores on the reading test. As a proof of its validity, the same measure collected during the reading comprehension test also predicted participants' Simon effect (incongruent - congruent response times) which is a common index of selective attention. This novel method allows researchers to detect attention shifts moments during reading without interrupting natural reading process.