A mixed filter algorithm for cognitive state estimation from simultaneously recorded continuous and binary measures of performance

Bayesian Inference of Hidden Cognitive Performance and Arousal States in Presence of Music

Goal: Poor arousal management may lead to reduced cognitive performance. Specifying a model and decoder to infer the cognitive arousal and performance contributes to arousal regulation via non-invasive actuators such as music. Methods: We employ a Bayesian filtering approach within an expectation-maximization framework to track the hidden states during the [Formula: see text]-back task in the presence of calming and exciting music. We decode the arousal and performance states from the skin conductance and behavioral signals, respectively. We derive an arousal-performance model based on the Yerkes-Dodson law. We design a performance-based arousal decoder by considering the corresponding performance and skin conductance as the observation. Results: The quantified arousal and performance are presented. The existence of Yerkes-Dodson law can be interpreted from the arousal-performance relationship. Findings display higher matrices of performance within the exciting music. Conclusions: The performance-based arousal decoder has a better agreement with the Yerkes-Dodson law. Our study can be implemented in designing non-invasive closed-loop systems.

Inferring cognitive state underlying conflict choices in verbal Stroop task using heterogeneous input discriminative-generative decoder model

Objective. The subthalamic nucleus (STN) of the basal ganglia interacts with the medial prefrontal cortex (mPFC) and shapes a control loop, specifically when the brain receives contradictory information from either different sensory systems or conflicting information from sensory inputs and prior knowledge that developed in the brain. Experimental studies demonstrated that significant increases in theta activities (2-8 Hz) in both the STN and mPFC as well as increased phase synchronization between mPFC and STN are prominent features of conflict processing. While these neural features reflect the importance of STN-mPFC circuitry in conflict processing, a low-dimensional representation of the mPFC-STN interaction referred to as a cognitive state, that links neural activities generated by these sub-regions to behavioral signals (e.g. the response time), remains to be identified.Approach. Here, we propose a new model, namely, the heterogeneous input discriminative-generative decoder (HI-DGD) model, to infer a cognitive state underlying decision-making based on neural activities (STN and mPFC) and behavioral signals (individuals' response time) recorded in ten Parkinson's disease (PD) patients while they performed a Stroop task. PD patients may have conflict processing which is quantitatively (may be qualitative in some) different from healthy populations.Main results. Using extensive synthetic and experimental data, we showed that the HI-DGD model can diffuse information from neural and behavioral data simultaneously and estimate cognitive states underlying conflict and non-conflict trials significantly better than traditional methods. Additionally, the HI-DGD model identified which neural features made significant contributions to conflict and non-conflict choices. Interestingly, the estimated features match well with those reported in experimental studies.Significance. Finally, we highlight the capability of the HI-DGD model in estimating a cognitive state from a single trial of observation, which makes it appropriate to be utilized in closed-loop neuromodulation systems.

Closed-Loop Tracking and Regulation of Emotional Valence State From Facial Electromyogram Measurements

Affective studies provide essential insights to address emotion recognition and tracking. In traditional open-loop structures, a lack of knowledge about the internal emotional state makes the system incapable of adjusting stimuli parameters and automatically responding to changes in the brain. To address this issue, we propose to use facial electromyogram measurements as biomarkers to infer the internal hidden brain state as feedback to close the loop. In this research, we develop a systematic way to track and control emotional valence, which codes emotions as being pleasant or obstructive. Hence, we conduct a simulation study by modeling and tracking the subject's emotional valence dynamics using state-space approaches. We employ Bayesian filtering to estimate the person-specific model parameters along with the hidden valence state, using continuous and binary features extracted from experimental electromyogram measurements. Moreover, we utilize a mixed-filter estimator to infer the secluded brain state in a real-time simulation environment. We close the loop with a fuzzy logic controller in two categories of regulation: inhibition and excitation. By designing a control action, we aim to automatically reflect any required adjustments within the simulation and reach the desired emotional state levels. Final results demonstrate that, by making use of physiological data, the proposed controller could effectively regulate the estimated valence state. Ultimately, we envision future outcomes of this research to support alternative forms of self-therapy by using wearable machine interface architectures capable of mitigating periods of pervasive emotions and maintaining daily well-being and welfare.

Direct Discriminative Decoder Models for Analysis of High-Dimensional Dynamical Neural Data

With the accelerated development of neural recording technology over the past few decades, research in integrative neuroscience has become increasingly reliant on data analysis methods that are scalable to high-dimensional recordings and computationally tractable. Latent process models have shown promising results in estimating the dynamics of cognitive processes using individual models for each neuron's receptive field. However, scaling these models to work on high-dimensional neural recordings remains challenging. Not only is it impractical to build receptive field models for individual neurons of a large neural population, but most neural data analyses based on individual receptive field models discard the local history of neural activity, which has been shown to be critical in the accurate inference of the underlying cognitive processes. Here, we propose a novel, scalable latent process model that can directly estimate cognitive process dynamics without requiring precise receptive field models of individual neurons or brain nodes. We call this the direct discriminative decoder (DDD) model. The DDD model consists of (1) a discriminative process that characterizes the conditional distribution of the signal to be estimated, or state, as a function of both the current neural activity and its local history, and (2) a state transition model that characterizes the evolution of the state over a longer time period. While this modeling framework inherits advantages of existing latent process modeling methods, its computational cost is tractable. More important, the solution can incorporate any information from the history of neural activity at any timescale in computing the estimate of the state process. There are many choices in building the discriminative process, including deep neural networks or gaussian processes, which adds to the flexibility of the framework. We argue that these attributes of the proposed methodology, along with its applicability to different modalities of neural data, make it a powerful tool for high-dimensional neural data analysis. We also introduce an extension of these methods, called the discriminative-generative decoder (DGD). The DGD includes both discriminative and generative processes in characterizing observed data. As a result, we can combine physiological correlates like behavior with neural data to better estimate underlying cognitive processes. We illustrate the methods, including steps for inference and model identification, and demonstrate applications to multiple data analysis problems with high-dimensional neural recordings. The modeling results demonstrate the computational and modeling advantages of the DDD and DGD methods.

A Mixed Filtering Approach for Real-Time Seizure State Tracking Using Multi-Channel Electroencephalography Data

Real-time continuous tracking of seizure state is necessary to develop feedback neuromodulation therapy that can prevent or terminate a seizure early. Due to its high temporal resolution, high scalp coverage, and non-invasive applicability, electroencephalography (EEG) is a good candidate for seizure tracking. In this research, we make multiple seizure state estimations using a mixed-filter and multiple channels found over the entire sensor space; then by applying a Kalman filter, we produce a single seizure state estimation made up of these individual estimations. Using a modified wrapper feature selection, we determine two optimal features of mixed data type, one continuous and one binary analyzing all available channels. These features are used in a state-space framework to model the continuous hidden seizure state. Expectation maximization is performed offline on the training and validation data sets to estimate unknown parameters. The seizure state estimation process is performed for multiple channels, and the seizure state estimation is derived using a square-root Kalman filter. A second expectation maximization step is utilized to estimate the unknown square-root Kalman filter parameters. This method is tested in a real-time applicable way for seizure state estimation. Applying this approach, we obtain a single seizure state estimation with quantitative information about the likelihood of a seizure occurring, which we call seizure probability. Our results on the experimental data (CHB-MIT EEG database) validate the proposed estimation method and we achieve an average accuracy, sensitivity, and specificity of 92.7%, 92.8%, and 93.4%, respectively. The potential applications of this seizure estimation model are for closed-loop neuromodulation and long-term quantitative analysis of seizure treatment efficacy.

Continuous Prediction of Cognitive State Using A Marked-Point Process Modeling Framework

Behavioral outcomes in many cognitive tasks are often recorded in a trial structure at discrete times. To adapt to this structure, neural encoder and decoder models have been built to take into account the trial organization to characterize the connection between brain dynamics and behavior, e.g. through latent dynamical models. The challenge of these models is that they are limited to discrete trial times while neural data is continuous. Here, we propose a marked-point process framework to characterize multivariate behavioral outcomes recorded during a trial-structured cognitive task, to build an estimation of cognitive state at a fine time resolution. We propose a state-space marked-point process modeling framework to characterize the relationship between observed behavior and underlying dynamical cognitive processes. We define the framework for a class of behavioral readouts by a response time and a discrete mark signifying an observed binary decision, and develop the state estimation and system identification steps. We define the filter and smoother for the marked-point process observation and develop an EM algorithm to estimate the model's free parameters. We demonstrate this modeling approach in a behavioral readout captured while participants perform an emotional conflict resolution task (ECR) and show that we can estimate underlying cognitive processes at a fine temporal resolution beyond the trial by trial approach.

A Bayesian Filtering Approach for Tracking Arousal From Binary and Continuous Skin Conductance Features

OBJECTIVE

Neuroanatomical structures within the cortical and sub-cortical brain regions process emotion and cause subsequent variations in signals such as skin conductance and electrocardiography. The signals often encode information in their continuous-valued amplitudes or waves as well as in their underlying impulsive events. We propose to track psychological arousal from this hybrid source of skin conductance information.

METHODS

We present a point process state-space method in tandem with Bayesian filtering for determining a continuous-valued arousal state from skin conductance measurements. To perform state estimation, we relate arousal to binary- and continuous-valued observations derived from the phasic and tonic parts of a skin conductance signal, and recover model parameters using expectation-maximization. We evaluate our model on both synthetic and two different experimental data sets. Stress was artificially induced in the first experimental data set and the second comprised of a fear conditioning experiment.

RESULTS

Results on the first data set indicate high levels of arousal during exposure to cognitive stress and low arousal during relaxation. Plausible results are also obtained in the fear conditioning data set consistent with previous skin conductance studies in similar experimental contexts.

CONCLUSION

The state-space approach-which does not rely on external classification labels-is able to continuously track an arousal level from skin conductance features.

SIGNIFICANCE

The method is a promising arousal estimation scheme utilizing only skin conductance. The approach could find applications in wearable monitoring and the study of neuropsychiatric conditions such as post-traumatic stress disorder.

Decoding Hidden Cognitive States From Behavior and Physiology Using a Bayesian Approach

Cognitive processes, such as learning and cognitive flexibility, are both difficult to measure and to sample continuously using objective tools because cognitive processes arise from distributed, high-dimensional neural activity. For both research and clinical applications, that dimensionality must be reduced. To reduce dimensionality and measure underlying cognitive processes, we propose a modeling framework in which a cognitive process is defined as a low-dimensional dynamical latent variable—called a cognitive state, which links high-dimensional neural recordings and multidimensional behavioral readouts. This framework allows us to decompose the hard problem of modeling the relationship between neural and behavioral data into separable encoding-decoding approaches. We first use a state-space modeling framework, the behavioral decoder, to articulate the relationship between an objective behavioral readout (e.g., response times) and cognitive state. The second step, the neural encoder, involves using a generalized linear model (GLM) to identify the relationship between the cognitive state and neural signals, such as local field potential (LFP). We then use the neural encoder model and a Bayesian filter to estimate cognitive state using neural data (LFP power) to generate the neural decoder. We provide goodness-of-fit analysis and model selection criteria in support of the encoding-decoding result. We apply this framework to estimate an underlying cognitive state from neural data in human participants ([Formula: see text]) performing a cognitive conflict task. We successfully estimated the cognitive state within the 95% confidence intervals of that estimated using behavior readout for an average of 90% of task trials across participants. In contrast to previous encoder-decoder models, our proposed modeling framework incorporates LFP spectral power to encode and decode a cognitive state. The framework allowed us to capture the temporal evolution of the underlying cognitive processes, which could be key to the development of closed-loop experiments and treatments.

Multimodal Encoding of Novelty, Reward, and Learning in the Primate Nucleus Basalis of Meynert

Associative learning is crucial for daily function, involving a complex network of brain regions. One region, the nucleus basalis of Meynert (NBM), is a highly interconnected, largely cholinergic structure implicated in multiple aspects of learning. We show that single neurons in the NBM of nonhuman primates (NHPs; n = 2 males; Macaca mulatta) encode learning a new association through spike rate modulation. However, the power of low-frequency local field potential (LFP) oscillations decreases in response to novel, not-yet-learned stimuli but then increase as learning progresses. Both NBM and the dorsolateral prefrontal cortex encode confidence in novel associations by increasing low- and high-frequency LFP power in anticipation of expected rewards. Finally, NBM high-frequency power dynamics are anticorrelated with spike rate modulations. Therefore, novelty, learning, and reward anticipation are separately encoded through differentiable NBM signals. By signaling both the need to learn and confidence in newly acquired associations, NBM may play a key role in coordinating cortical activity throughout the learning process.SIGNIFICANCE STATEMENT Degradation of cells in a key brain region, the nucleus basalis of Meynert (NBM), correlates with Alzheimer's disease and Parkinson's disease progression. To better understand the role of this brain structure in learning and memory, we examined neural activity in the NBM in behaving nonhuman primates while they performed a learning and memory task. We found that single neurons in NBM encoded both salience and an early learning, or cognitive state, whereas populations of neurons in the NBM and prefrontal cortex encode learned state and reward anticipation. The NBM may thus encode multiple stages of learning. These multimodal signals might be leveraged in future studies to develop neural stimulation to facilitate different stages of learning and memory.

Estimating Dynamic Signals From Trial Data With Censored Values

Censored data occur commonly in trial-structured behavioral experiments and many other forms of longitudinal data. They can lead to severe bias and reduction of statistical power in subsequent analyses. Principled approaches for dealing with censored data, such as data imputation and methods based on the complete data's likelihood, work well for estimating fixed features of statistical models but have not been extended to dynamic measures, such as serial estimates of an underlying latent variable over time. Here we propose an approach to the censored-data problem for dynamic behavioral signals. We developed a state-space modeling framework with a censored observation process at the trial timescale. We then developed a filter algorithm to compute the posterior distribution of the state process using the available data. We showed that special cases of this framework can incorporate the three most common approaches to censored observations: ignoring trials with censored data, imputing the censored data values, or using the full information available in the data likelihood. Finally, we derived a computationally efficient approximate Gaussian filter that is similar in structure to a Kalman filter, but that efficiently accounts for censored data. We compared the performances of these methods in a simulation study and provide recommendations of approaches to use, based on the expected amount of censored data in an experiment. These new techniques can broadly be applied in many research domains in which censored data interfere with estimation, including survival analysis and other clinical trial applications.

Tracking the sleep onset process: an empirical model of behavioral and physiological dynamics

The sleep onset process (SOP) is a dynamic process correlated with a multitude of behavioral and physiological markers. A principled analysis of the SOP can serve as a foundation for answering questions of fundamental importance in basic neuroscience and sleep medicine. Unfortunately, current methods for analyzing the SOP fail to account for the overwhelming evidence that the wake/sleep transition is governed by continuous, dynamic physiological processes. Instead, current practices coarsely discretize sleep both in terms of state, where it is viewed as a binary (wake or sleep) process, and in time, where it is viewed as a single time point derived from subjectively scored stages in 30-second epochs, effectively eliminating SOP dynamics from the analysis. These methods also fail to integrate information from both behavioral and physiological data. It is thus imperative to resolve the mismatch between the physiological evidence and analysis methodologies. In this paper, we develop a statistically and physiologically principled dynamic framework and empirical SOP model, combining simultaneously-recorded physiological measurements with behavioral data from a novel breathing task requiring no arousing external sensory stimuli. We fit the model using data from healthy subjects, and estimate the instantaneous probability that a subject is awake during the SOP. The model successfully tracked physiological and behavioral dynamics for individual nights, and significantly outperformed the instantaneous transition models implicit in clinical definitions of sleep onset. Our framework also provides a principled means for cross-subject data alignment as a function of wake probability, allowing us to characterize and compare SOP dynamics across different populations. This analysis enabled us to quantitatively compare the EEG of subjects showing reduced alpha power with the remaining subjects at identical response probabilities. Thus, by incorporating both physiological and behavioral dynamics into our model framework, the dynamics of our analyses can finally match those observed during the SOP.

Dissociating hippocampal and striatal contributions to sequential prediction learning

Behavior may be generated on the basis of many different kinds of learned contingencies. For instance, responses could be guided by the direct association between a stimulus and response, or by sequential stimulus-stimulus relationships (as in model-based reinforcement learning or goal-directed actions). However, the neural architecture underlying sequential predictive learning is not well understood, in part because it is difficult to isolate its effect on choice behavior. To track such learning more directly, we examined reaction times (RTs) in a probabilistic sequential picture identification task in healthy individuals. We used computational learning models to isolate trial-by-trial effects of two distinct learning processes in behavior, and used these as signatures to analyse the separate neural substrates of each process. RTs were best explained via the combination of two delta rule learning processes with different learning rates. To examine neural manifestations of these learning processes, we used functional magnetic resonance imaging to seek correlates of time-series related to expectancy or surprise. We observed such correlates in two regions, hippocampus and striatum. By estimating the learning rates best explaining each signal, we verified that they were uniquely associated with one of the two distinct processes identified behaviorally. These differential correlates suggest that complementary anticipatory functions drive each region's effect on behavior. Our results provide novel insights as to the quantitative computational distinctions between medial temporal and basal ganglia learning networks and enable experiments that exploit trial-by-trial measurement of the unique contributions of both hippocampus and striatum to response behavior.

Characterizing learning by simultaneous analysis of continuous and binary measures of performance

Continuous observations, such as reaction and run times, and binary observations, such as correct/incorrect responses, are recorded routinely in behavioral learning experiments. Although both types of performance measures are often recorded simultaneously, the two have not been used in combination to evaluate learning. We present a state-space model of learning in which the observation process has simultaneously recorded continuous and binary measures of performance. We use these performance measures simultaneously to estimate the model parameters and the unobserved cognitive state process by maximum likelihood using an approximate expectation maximization (EM) algorithm. We introduce the concept of a reaction-time curve and reformulate our previous definitions of the learning curve, the ideal observer curve, the learning trial and between-trial comparisons of performance in terms of the new model. We illustrate the properties of the new model in an analysis of a simulated learning experiment. In the simulated data analysis, simultaneous use of the two measures of performance provided more credible and accurate estimates of the learning than either measure analyzed separately. We also analyze two actual learning experiments in which the performance of rats and of monkeys was tracked across trials by simultaneously recorded reaction and run times and the correct and incorrect responses. In the analysis of the actual experiments, our algorithm gave a straightforward, efficient way to characterize learning by combining continuous and binary measures of performance. This analysis paradigm has implications for characterizing learning and for the more general problem of combining different data types to characterize the properties of a neural system.

Discrete- and continuous-time probabilistic models and algorithms for inferring neuronal UP and DOWN states

UP and DOWN states, the periodic fluctuations between increased and decreased spiking activity of a neuronal population, are a fundamental feature of cortical circuits. Understanding UP-DOWN state dynamics is important for understanding how these circuits represent and transmit information in the brain. To date, limited work has been done on characterizing the stochastic properties of UP-DOWN state dynamics. We present a set of Markov and semi-Markov discrete- and continuous-time probability models for estimating UP and DOWN states from multiunit neural spiking activity. We model multiunit neural spiking activity as a stochastic point process, modulated by the hidden (UP and DOWN) states and the ensemble spiking history. We estimate jointly the hidden states and the model parameters by maximum likelihood using an expectation-maximization (EM) algorithm and a Monte Carlo EM algorithm that uses reversible-jump Markov chain Monte Carlo sampling in the E-step. We apply our models and algorithms in the analysis of both simulated multiunit spiking activity and actual multi- unit spiking activity recorded from primary somatosensory cortex in a behaving rat during slow-wave sleep. Our approach provides a statistical characterization of UP-DOWN state dynamics that can serve as a basis for verifying and refining mechanistic descriptions of this process.