Jason F. Smith

Identification and validation of effective connectivity networks in functional magnetic resonance imaging using switching linear dynamic systems.

Dynamic connectivity networks identify directed interregional interactions between modeled brain regions in neuroimaging. However, problems arise when the regions involved in a task and their interconnections are not fully known a priori. Objective measures of model adequacy are necessary to validate such models. We present a connectivity formalism, the Switching Linear Dynamic System (SLDS), that is capable of identifying both Granger-Geweke and instantaneous connectivity that vary according to experimental conditions. SLDS explicitly models the task condition as a Markov random variable. The series of task conditions can be estimated from new data given an identified model providing a means to validate connectivity patterns. We use SLDS to model functional magnetic resonance imaging data from five regions during a finger alternation task. Using interregional connectivity alone, the identified model predicted the task condition vector from a different subject with a different task ordering with high accuracy. In addition, important regions excluded from a model can be identified by augmenting the model state space. A motor task model excluding primary motor cortices was augmented with a new neural state constrained by its connectivity with the included regions. The augmented variable time series, convolved with a hemodynamic kernel, was compared to all brain voxels. The right primary motor cortex was identified as the best region to add to the model. Our results suggest that the SLDS model framework is an effective means to address several problems with modeling connectivity including measuring overall model adequacy and identifying important regions missing from models.

Network analysis of single-subject fMRI during a finger opposition task

The analysis of functional magnetic resonance imaging (fMRI) data has typically relied on univariate methods to identify areas of brain activity related to cognitive and behavioral task performance. We investigated the ability of multivariate network analysis using a modified form of principal component analysis, the Scaled Subprofile Model (SSM), applied to single-subject fMRI data to identify patterns of interactions among brain regions over time during an anatomically well-characterized simple motor task. We hypothesized that each subject would exhibit correlated patterns of brain activation in several regions known to participate in the regulation of movement including the contralateral motor cortex and the ipsilateral cerebellum. EPI BOLD images were acquired in six healthy participants as they performed a visually and auditorally paced finger opposition task. SSM analysis was applied to the fMR time series on a single-subject basis. Linear combinations of the major principal components that predicted the expected hemodynamic response to the order of experimental conditions were identified for each participant. These combinations of SSM patterns were highly associated with the expected hemodynamic response, an indicator of local neuronal activity, in each participant (0.84 </= R(2) </= 0.97, all Ps < 0.0001). As predicted, the combined pattern in each subject was characterized most prominently by relatively increased activations in contralateral sensorimotor cortex and ipsilateral cerebellum. Additionally, all subjects showed areas of relatively decreased activation in the ipsilateral sensorimotor cortex and contralateral cerebellum. The application of network analysis methods, such as SSM, to single-subject fMRI data can identify patterns of task-specific, functionally interacting brain areas in individual subjects. This approach may help identify individual differences in the task-related functional connectivity, track changes in task-related patterns of activity within or between fMRI sessions, and provide a method to identify individual differences in response to treatment.

Discrimination task reveals differences in neural bases of tinnitus and hearing impairment

We investigated auditory perception and cognitive processing in individuals with chronic tinnitus or hearing loss using functional magnetic resonance imaging (fMRI). Our participants belonged to one of three groups: bilateral hearing loss and tinnitus (TIN), bilateral hearing loss without tinnitus (HL), and normal hearing without tinnitus (NH). We employed pure tones and frequency-modulated sweeps as stimuli in two tasks: passive listening and active discrimination. All subjects had normal hearing through 2 kHz and all stimuli were low-pass filtered at 2 kHz so that all participants could hear them equally well. Performance was similar among all three groups for the discrimination task. In all participants, a distributed set of brain regions including the primary and non-primary auditory cortices showed greater response for both tasks compared to rest. Comparing the groups directly, we found decreased activation in the parietal and frontal lobes in the participants with tinnitus compared to the HL group and decreased response in the frontal lobes relative to the NH group. Additionally, the HL subjects exhibited increased response in the anterior cingulate relative to the NH group. Our results suggest that a differential engagement of a putative auditory attention and short-term memory network, comprising regions in the frontal, parietal and temporal cortices and the anterior cingulate, may represent a key difference in the neural bases of chronic tinnitus accompanied by hearing loss relative to hearing loss alone.

Effective Connectivity Modeling for fMRI: Six Issues and Possible Solutions Using Linear Dynamic Systems

Analysis of directionally specific or causal interactions between regions in functional magnetic resonance imaging (fMRI) data has proliferated. Here we identify six issues with existing effective connectivity methods that need to be addressed. The issues are discussed within the framework of linear dynamic systems for fMRI (LDSf). The first concerns the use of deterministic models to identify inter-regional effective connectivity. We show that deterministic dynamics are incapable of identifying the trial-to-trial variability typically investigated as the marker of connectivity while stochastic models can capture this variability. The second concerns the simplistic (constant) connectivity modeled by most methods. Connectivity parameters of the LDSf model can vary at the same timescale as the input data. Further, extending LDSf to mixtures of multiple models provides more robust connectivity variation. The third concerns the correct identification of the network itself including the number and anatomical origin of the network nodes. Augmentation of the LDSf state space can identify additional nodes of a network. The fourth concerns the locus of the signal used as a “node” in a network. A novel extension LDSf incorporating sparse canonical correlations can select most relevant voxels from an anatomically defined region based on connectivity. The fifth concerns connection interpretation. Individual parameter differences have received most attention. We present alternative network descriptors of connectivity changes which consider the whole network. The sixth concerns the temporal resolution of fMRI data relative to the timescale of the inter-regional interactions in the brain. LDSf includes an “instantaneous” connection term to capture connectivity occurring at timescales faster than the data resolution. The LDS framework can also be extended to statistically combine fMRI and EEG data. The LDSf framework is a promising foundation for effective connectivity analysis.

The functional overlap of executive control and language processing in bilinguals.

The need to control multiple languages is thought to require domain-general executive control (EC) in bilinguals such that the EC and language systems become interdependent. However, there has been no systematic investigation into how and where EC and language processes overlap in the bilingual brain. If the concurrent recruitment of EC during bilingual language processing is domain-general and extends to non-linguistic EC, we hypothesize that regions commonly involvement in language processing, linguistic EC, and non-linguistic EC may be selectively altered in bilinguals compared to monolinguals. A conjunction of functional magnetic resonance imaging (fMRI) data from a flanker task with linguistic and nonlinguistic distractors and a semantic categorization task showed functional overlap in the left inferior frontal gyrus (LIFG) in bilinguals, whereas no overlap occurred in monolinguals. This research therefore identifies a neural locus of functional overlap of language and EC in the bilingual brain.

Imaging systems level consolidation of novel associate memories: A longitudinal neuroimaging study

Previously, a standard theory of systems level memory consolidation was developed to describe how memory recall becomes independent of the medial temporal memory system. More recently, an extended consolidation theory was proposed that predicts seven changes in regional neural activity and inter-regional functional connectivity. Using longitudinal event-related functional magnetic resonance imaging of an associate memory task, we simultaneously tested all predictions and additionally tested for consolidation-related changes in recall of associate memories at a sub-trial temporal resolution, analyzing cue, delay and target periods of each trial separately. Results consistent with the theoretical predictions were observed though two inconsistent results were also obtained. In particular, while medial temporal recall related delay period activity decreased with consolidation as predicted, visual cue activity increased for consolidated memories. Though the extended theory of memory consolidation is largely supported by our study, these results suggest that the extended theory needs further refinement and the medial temporal memory system has multiple, temporally distinct roles in associate memory recall. Neuroimaging analysis at a sub-trial temporal resolution, as used here, may further clarify the role of the hippocampal complex in memory consolidation.

A link between neuroscience and informatics: large-scale modeling of memory processes.

Utilizing advances in functional neuroimaging and computational neural modeling, neuroscientists have increasingly sought to investigate how distributed networks, composed of functionally defined subregions, combine to produce cognition. Large-scale, biologically realistic neural models, which integrate data from cellular, regional, whole brain, and behavioral sources, delineate specific hypotheses about how these interacting neural populations might carry out high-level cognitive tasks. In this review, we discuss neuroimaging, neural modeling, and the utility of large-scale biologically realistic models using modeling of short-term memory as an example. We present a sketch of the data regarding the neural basis of short-term memory from non-human electrophysiological, computational and neuroimaging perspectives, highlighting the multiple interacting brain regions believed to be involved. Through a review of several efforts, including our own, to combine neural modeling and neuroimaging data, we argue that large scale neural models provide specific advantages in understanding the distributed networks underlying cognition and behavior.

Separating lexical-semantic access from other mnemonic processes in picture-name verification.

We present a novel paradigm to identify shared and unique brain regions underlying non-semantic, non-phonological, abstract, audio-visual (AV) memory vs. naming using a longitudinal functional magnetic resonance imaging experiment. Participants were trained to associate novel AV stimulus pairs containing hidden linguistic content. Half of the stimulus pairs were distorted images of animals and sine-wave speech versions of the animals name. Images and sounds were distorted in such a way as to make their linguistic content easily recognizable only after being made aware of its existence. Memory for the pairings was tested by presenting an AV pair and asking participants to verify if the two stimuli formed a learned pairing. After memory testing, the hidden linguistic content was revealed and participants were tested again on their recollection of the pairings in this linguistically informed state. Once informed, the AV verification task could be performed by naming the picture. There was substantial overlap between the regions involved in recognition of non-linguistic sensory memory and naming, suggesting a strong relation between them. Contrasts between sessions identified left angular gyrus and middle temporal gyrus as key additional players in the naming network. Left inferior frontal regions participated in both naming and non-linguistic AV memory suggesting the region is responsible for AV memory independent of phonological content contrary to previous proposals. Functional connectivity between angular gyrus and left inferior frontal gyrus and left middle temporal gyrus increased when performing the AV task as naming. The results are consistent with the hypothesis that, at the spatial resolution of fMRI, the regions that facilitate non-linguistic AV associations are a subset of those that facilitate naming though reorganized into distinct networks.

Identifying effective connectivity parameters in simulated fMRI: a direct comparison of switching linear dynamic system, stochastic dynamic causal, and multivariate autoregressive models

The number and variety of connectivity estimation methods is likely to continue to grow over the coming decade. Comparisons between methods are necessary to prune this growth to only the most accurate and robust methods. However, the nature of connectivity is elusive with different methods potentially attempting to identify different aspects of connectivity. Commonalities of connectivity definitions across methods upon which base direct comparisons can be difficult to derive. Here, we explicitly define “effective connectivity” using a common set of observation and state equations that are appropriate for three connectivity methods: dynamic causal modeling (DCM), multivariate autoregressive modeling (MAR), and switching linear dynamic systems for fMRI (sLDSf). In addition while deriving this set, we show how many other popular functional and effective connectivity methods are actually simplifications of these equations. We discuss implications of these connections for the practice of using one method to simulate data for another method. After mathematically connecting the three effective connectivity methods, simulated fMRI data with varying numbers of regions and task conditions is generated from the common equation. This simulated data explicitly contains the type of the connectivity that the three models were intended to identify. Each method is applied to the simulated data sets and the accuracy of parameter identification is analyzed. All methods perform above chance levels at identifying correct connectivity parameters. The sLDSf method was superior in parameter estimation accuracy to both DCM and MAR for all types of comparisons.

Temporal microstructure of cortical networks (TMCN) underlying task-related differences

Neuro-electromagnetic recording techniques (EEG, MEG, iEEG) provide high temporal resolution data to study the dynamics of neurocognitive networks: large scale neural assemblies involved in task-specific information processing. How does a neurocognitive network reorganize spatiotemporally on the order of a few milliseconds to process specific aspects of the task? At what times do networks segregate for task processing, and at what time scales does integration of information occur via changes in functional connectivity? Here, we propose a data analysis framework-Temporal microstructure of cortical networks (TMCN)-that answers these questions for EEG/MEG recordings in the signal space. Method validation is established on simulated MEG data from a delayed-match to-sample (DMS) task. We then provide an example application on MEG recordings during a paired associate task (modified from the simpler DMS paradigm) designed to study modality specific long term memory recall. Our analysis identified the times at which network segregation occurs for processing the memory recall of an auditory object paired to a visual stimulus (visual-auditory) in comparison to an analogous visual-visual pair. Across all subjects, onset times for first network divergence appeared within a range of 0.08-0.47 s after initial visual stimulus onset. This indicates that visual-visual and visual auditory memory recollection involves equivalent network components without any additional recruitment during an initial period of the sensory processing stage which is then followed by recruitment of additional network components for modality specific memory recollection. Therefore, we propose TMCN as a viable computational tool for extracting network timing in various cognitive tasks.