Neural modeling and functional brain imaging: an overview

Method matters: an empirical study of impact in cognitive neuroscience

A major thrust of cognitive neuroscience is the elucidation of structure-function relationships in the human brain. Over the last several years, functional neuroimaging has risen in prominence relative to the lesion studies that formed the historical core of work in this field. These two methods have different strengths and weaknesses. Among these is a crucial difference in the nature of evidence each can provide. Lesion studies can provide evidence for necessity claims, whereas functional neuroimaging studies do not. We hypothesized that lesion studies will continue to have greater scientific impact even as the relative proportion of such studies in the cognitive neuroscience literature declines. Using methods drawn from systematic literature review, we identified a set of original cognitive neuroscience articles that employed either functional imaging or lesion techniques, published at one of two time points in the 1990s, and assessed the effect of the method used on each article's impact across the decade. Functional neuroimaging studies were cited three times more often than lesion studies throughout the time span we examined. This effect was in large part due to differences in the influence of the journals publishing the two methods; functional neuroimaging studies appeared disproportionately more often in higher impact journals. There were also differences in the degree to which articles using one method cited articles using the other method. Functional neuroimaging articles were less likely to include such cross-method citations.

Combined event-related fMRI and intracerebral ERP study of an auditory oddball task

Event-related fMRI (efMRI) has been repeatedly used to seek the neural sources of endogenous event-related potentials (ERP). However, significant discrepancies exist between the efMRI data and the results of previously published intracranial ERP studies of oddball task. To evaluate the capacity of efMRI to define the sources of the P3 component of ERP within the human brain, both efMRI and intracerebral ERP recordings were performed in eight patients with intractable epilepsy (five males and three females) during their preoperative invasive video-EEG monitoring. An identical auditory oddball task with frequent and target stimuli was completed in two sessions. A total of 606 intracerebral sites were electrophysiologically investigated by means of depth electrodes. In accordance with the finding of multiple intracerebral generators of P3 potential, the target stimuli evoked MRI signal increase in multiple brain regions. However, regions with evident hemodynamic and electrophysiological responses overlapped only partially. P3 generators were always found within hemodynamic-active sites, if these sites were investigated by means of depth electrodes. On the other hand, unequivocal local sources of P3 potential were apparently also located outside the regions with a significant hemodynamic response (typically in mesiotemporal regions). Both methods should thus be viewed as mutually complementary in investigations of the spatial distribution of cortical and subcortical activation during oddball task.

Investigating the neural basis for functional and effective connectivity. Application to fMRI

Viewing cognitive functions as mediated by networks has begun to play a central role in interpreting neuroscientific data, and studies evaluating interregional functional and effective connectivity have become staples of the neuroimaging literature. The neurobiological substrates of functional and effective connectivity are, however, uncertain. We have constructed neurobiologically realistic models for visual and auditory object processing with multiple interconnected brain regions that perform delayed match-to-sample (DMS) tasks. We used these models to investigate how neurobiological parameters affect the interregional functional connectivity between functional magnetic resonance imaging (fMRI) time-series. Variability is included in the models as subject-to-subject differences in the strengths of anatomical connections, scan-to-scan changes in the level of attention, and trial-to-trial interactions with non-specific neurons processing noise stimuli. We find that time-series correlations between integrated synaptic activities between the anterior temporal and the prefrontal cortex were larger during the DMS task than during a control task. These results were less clear when the integrated synaptic activity was haemodynamically convolved to generate simulated fMRI activity. As the strength of the model anatomical connectivity between temporal and frontal cortex was weakened, so too was the strength of the corresponding functional connectivity. These results provide a partial validation for using fMRI functional connectivity to assess brain interregional relations.

Integrating neuroscientific data across spatiotemporal scales

A major challenge confronting neuroscientists is associated with the multiple spatial and temporal scales of investigation of neural structure and function. I shall discuss the use of computational neural modeling as one method to bridge some of the different spatial and temporal levels. This approach will be illustrated using large-scale, neurobiologically realistic network models of auditory and visual pattern recognition that relate neuronal dynamics to fMRI data. It will be demonstrated that the models are capable of exhibiting the salient features of both electrophysiological neuronal activities and fMRI values that are in agreement with empirically observed data.

Models of Brain Function in Neuroimaging

Inferences about brain function, using neuroimaging data, rest on models of how the data were caused. These models can be quite diverse, ranging from conceptual models of functional anatomy to nonlinear mathematical models of hemodynamics. However, they all have to be internally consistent because they model the same thing. This consistency encompasses many levels of description and places constraints on the statistical models, adopted for data analysis, and the experimental designs they embody. The aim of this review is to introduce the key models used in imaging neuroscience and how they relate to each other. We start with anatomical models of functional brain architectures, which motivate some of the fundaments of neuroimaging. We then turn to basic statistical models (e.g., the general linear model) used for making classical and Bayesian inferences about where neuronal responses are expressed. By incorporating biophysical constraints, these basic models can be finessed and, in a dynamic setting, rendered causal. This allows us to infer how interactions among brain regions are mediated.

Fine Modulation in Network Activation during Motor Execution and Motor Imagery

Motor imagery, the 'mental rehearsal of motor acts without overt movements', involves either a visual representation (visual imagery, VI) or mental simulation of movement, associated with a kinesthetic feeling (kinetic imagery, KI). Previous brain imaging work suggests that patterns of brain activation differ when comparing execution (E) with either type of imagery but the functional connectivity of the participating networks has not been studied. Using functional magnetic resonance imaging (fMRI) and structural equation modeling, this study elucidates the inter-relationships among the relevant areas for each of the three motor behaviors. Our results suggest that networks underlying these behaviors are not identical, despite the extensive overlap between E and KI. Inputs to M1, which are facilitatory during E, have the opposite effect during KI, suggesting a physiological mechanism whereby the system prevents overt movements. Finally, this study highlights the role of the connection of superior parietal lobule to the supplementary motor area in both types of motor imagery.

Relating fMRI and PET signals to neural activity by means of large-scale neural models

This article reviews the four ways by which large-scale, neurobiologically realistic modeling can be used in conjunction with functional neuroimaging data, especially that obtained by functional magnetic resonance imaging (fMRI) and positron emission tomography (PET), to help investigators understand the neural bases for sensorimotor and cognitive functions. The conceptually distinct purposes served are:(1) formulating and implementing specific hypotheses about how neuronal populations mediate a task, which will be illustrated using models of visual and auditory object processing; (2) determining how well an experimental design paradigm or analysis method works, which will be illustrated by examining event-related fMRI; (3) investigating the meaning in neural terms of macro-level concepts, which will be illustrated using functional connectivity; and (4) combining different types of macroscopic data with one another, which will be illustrated using transcranial magnetic stimulation (TMS) and PET.

Relating neuronal dynamics for auditory object processing to neuroimaging activity: a computational modeling and an fMRI study

We investigated the neural basis of auditory object processing in the cerebral cortex by combining neural modeling and functional neuroimaging. We developed a large-scale, neurobiologically realistic network model of auditory pattern recognition that relates the neuronal dynamics of cortical auditory processing of frequency modulated (FM) sweeps to functional neuroimaging data of the type obtained using PET and fMRI. Areas included in the model extend from primary auditory to prefrontal cortex. The electrical activities of the neuronal units of the model were constrained to agree with data from the neurophysiological literature regarding the perception of FM sweeps. We also conducted an fMRI experiment using stimuli and tasks similar to those used in our simulations. The integrated synaptic activity of the neuronal units in each region of the model, convolved with a hemodynamic response function, was used as a correlate of the simulated fMRI activity, and generally agreed with the experimentally observed fMRI data in the brain areas corresponding to the regions of the model. Our results demonstrate that the model is capable of exhibiting the salient features of both electrophysiological neuronal activities and fMRI values that are in agreement with empirically observed data. These findings provide support for our hypotheses concerning how auditory objects are processed by primate neocortex.

Limbic-frontal circuitry in major depression: a path modeling metanalysis

This paper reports the results of an across lab metanalysis of effective connectivity in major depression (MDD). Using FDG PET data and Structural Equation Modeling, a formal depression model was created to explicitly test current theories of limbic-cortical dysfunction in MDD and to characterize at the path level potential sources of baseline variability reported in this patient population. A 7-region model consisting of lateral prefrontal cortex (latF9), anterior thalamus (aTh), anterior cingulate (Cg24), subgenual cingulate (Cg25), orbital frontal cortex (OF11), hippocampus (Hc), and medial frontal cortex (mF10) was tested in scans of 119 depressed patients and 42 healthy control subjects acquired during three separate studies at two different institutions. A single model, based on previous theory and supported by anatomical connectivity literature, was stable for the three groups of depressed patients. Within the context of this model, path differences among groups as a function of treatment response characteristics were also identified. First, limbic-cortical connections (latF9-Cg25-OF11-Hc) differentiated drug treatment responders from nonresponders. Second, nonresponders showed additional abnormalities in limbic-subcortical pathways (aTh-Cg24-Cg25-OF11-Hc). Lastly, more limited limbic-cortical (Hc-latF9) and cortical-cortical (OF11-mF10) path differences differentiated responders to cognitive behavioral therapy (CBT) from responders to pharmacotherapy. We conclude that the creation of such models is a first step toward full characterization of the depression phenotype at the neural systems level, with implications for the future development of brain-based algorithms to determine optimal treatment selection for individual patients.

“What” and “Where” in Visual Working Memory: A Computational Neurodynamical Perspective for Integrating fMRI and Single-Neuron Data

Single-neuron recordings, functional magnetic resonance imaging (fMRI) data, and the effects of lesions indicate that the prefrontal cortex (PFC) is involved in some types of working memory and related cognitive processes. Based on these data, two different models of the topographical and functional organization of the PFC have been proposed: organization-by-stimulus-domain, and organization-by-process. In this article, we utilize an integrate-and-fire network to model both single-neuron and fMRI data on short-term memory in order to understand data obtained in topologically different parts of the PFC during working memory tasks. We explicitly model the mechanisms that underlie workingmemory-related activity during the execution of delay tasks that have a “what”-then-“where” design (with both object and spatial delayed responses within the same trial). The model contains different populations of neurons (as found experimentally) in attractor networks that respond in the delay period to the stimulus object, the stimulus position, and to combinations of both object and position information. The pools are arranged hierarchically and have global inhibition through inhibitory interneurons to implement competition. It is shown that a biasing attentional input to define the current relevant information (object or location) enables the system to select the correct neuronal populations during the delay period in what is a biased competition model of attention. The processes occurring at the AMPA and NMDA synapses are dynamically modeled in the integrate-and-fire implementation to produce realistic spiking dynamics. It is shown that the fMRI data characteristic of the dorsal PFC and linked to spatial processing and manipulation of items can be reproduced in the model by a high level of inhibition, whereas the fMRI data characteristic of the ventral PFC and linked to object processing can be produced by a lower level of inhibition, even though the network is itself topographically homogeneous with no spatial topology of the neurons. This article, thus, not only presents a model for how spatial versus object short-term memory could be implemented in the PFC, but also shows that the fMRI BOLD signal measured during such tasks from different parts of the PFC could reflect a higher level of inhibition dorsally, without this dorsal region necessarily being primarily spatial and the ventral region object-related.

Dopaminergic drug effects on physiological connectivity in a human cortico-striato-thalamic system

Cortico-striato-thalamic (CST) systems are anatomical substrates for many motor and executive functions and are implicated in diverse neuropsychiatric disorders. Electrophysiological studies in rats, monkeys and patients with Parkinson's disease have shown that power and coherence of low frequency oscillations in CST systems can be profoundly modulated by dopaminergic drugs. We combined functional MRI with correlational and path analyses to investigate functional and effective connectivity, respectively, of a prefronto-striato-thalamic system activated by object location learning in healthy elderly human subjects (n = 23; mean age = 72 years). Participants were scanned in a repeated measures, randomized, placebo-controlled design to measure modulation of physiological connectivity between CST regions following treatment with drugs which served both to decrease (sulpiride) and increase (methylphenidate) dopaminergic transmission, as well as non-dopaminergic treatments (diazepam and scopolamine) to examine non-specific effects. Functional connectivity of caudate nucleus was modulated specifically by dopaminergic drugs, with opposing effects of sulpiride and methylphenidate. The more salient effect of sulpiride was to increase functional connectivity between caudate and both thalamus and ventral midbrain. A path diagram based on prior knowledge of unidirectional anatomical projections between CST components was fitted satisfactorily to the observed inter-regional covariance matrix. The effect of sulpiride was defined more specifically in the context of this model as increased strength of effective connection from ventral midbrain to caudate nucleus. In short, we have demonstrated enhanced functional and effective connectivity of human caudate nucleus following sulpiride treatment, which is compatible both with the anatomy of ascending dopaminergic projections and with electrophysiological studies indicating abnormal coherent oscillations of CST neurons in parkinsonian states.

Neuroimaging studies of word and pseudoword reading: consistencies, inconsistencies, and limitations

Several functional neuroimaging studies have compared words and pseudowords to test different cognitive models of reading. There are difficulties with this approach, however, because cognitive models do not make clear-cut predictions at the neural level. Therefore, results can only be interpreted on the basis of prior knowledge of cognitive anatomy. Furthermore, studies comparing words and pseudowords have produced inconsistent results. The inconsistencies could reflect false-positive results due to the low statistical thresholds applied or confounds from nonlexical aspects of the stimuli. Alternatively, they may reflect true effects that are inconsistent across subjects; dependent on experimental parameters such as stimulus rate or duration; or not replicated across studies because of insufficient statistical power. In this fMRI study, we investigate consistent and inconsistent differences between word and pseudoword reading in 20 subjects, and distinguish between effects associated with increases and decreases in activity relative to fixation. In addition, the interaction of word type with stimulus duration is explored. We find that words and pseudowords activate the same set of regions relative to fixation, and within this system, there is greater activation for pseudowords than words in the left frontal operculum, left posterior inferior temporal gyrus, and the right cerebellum. The only effects of words relative to pseudowords consistent over subjects are due to decreases in activity for pseudowords relative to fixation; and there are no significant interactions between word type and stimulus duration. Finally, we observe inconsistent but highly significant effects of word type at the individual subject level. These results (i) illustrate that pseudowords place increased demands on areas that have previously been linked to lexical retrieval, and (ii) highlight the importance of including one or more baselines to qualify word type effects. Furthermore, (iii) they suggest that inconsistencies observed in the previous literature may result from effects arising from a small number of subjects only.

Diffusion-tensor imaging of cognitive performance

MR methods have for some years been used to assess cognitive performance. Recently, studies have shown that diffusion-tensor imaging (DTI), which provides noninvasive maps of microscopic structural information of oriented tissue in vivo, is finding utility in studies of cognition in the normal and abnormal aging population. These studies suggest that water proton nonrandom, anisotropic diffusion measured by DTI is highly sensitive to otherwise subtle disease processes not easily seen with conventional MRI tissue contrast mechanisms and raises new issues of the role of MR in assessing cognitive potential.

TOWARDS AN INTEGRATIVE THEORY OF COGNITION

A framework is outlined for connecting brain imaging activity with the underlying biophysical properties of neural networks, and their mechanisms of action and organizing principles. The main thrust of the framework is a dynamic theory of semantics based on functional integration of biophysical neural networks. It asserts that higher-brain function arises from both synaptic and extrasynaptic integration in the neuropil where information on environmental changes are represented dynamically through a discourse of semantics. Consequently, integrative neural modeling is shown to be an important methodology for analyzing the response activities of functional imaging studies in elucidating the relationship between brain and structure, function and behavior.

Functional anatomy of syntactic and semantic processing in language comprehension

A functional magnetic resonance imaging (fMRI) study was conducted to map syntactic and semantic processes onto the brain. Chinese-English bilingual subjects performed two experimental tasks: a syntactic plausibility judgment task in which they decided whether a viewed verb phrase was syntactically legal, and a semantic plausibility judgment task in which they decided whether a viewed phrase was semantically acceptable. A font size judgment task was used as baseline. It is found that a large-scale distributed neural network covering the left mid-inferior frontal and mid-superior temporal cortices was responsible for the processing of Chinese phrases. The right homologue areas of these left cortical sites were also active, although the brain activity was obviously left-lateralized. Unlike previous research with monolingual English speakers that showed that distinct brain regions mediate syntactic and semantic processing of English, the cortical sites contributing to syntactic analysis of Chinese phrases coincided with the cortical sites relevant to semantic analysis. Stronger brain activity, however, was seen in the left middle frontal cortex for syntactic processing (relative to semantic processing), whereas for semantic processing stronger cortical activations were shown in the left inferior prefrontal cortex and the left mid-superior temporal gyri. The overall pattern of results indicates that syntactic processing is less independent in reading Chinese. This is attributable to the linguistic nature of the Chinese language that semantics and syntax are not always clearly demarcated. Equally interesting, we discovered that when our bilingual subjects performed syntactic and semantic acceptability judgments of English phrases, they applied the cerebral systems underlying Chinese reading to the processing of English.

Neurobiological measures of human selective attention

Selective attention allows people to process some stimuli more thoroughly than others. This is partly under voluntary control, and partly determined by stimulus salience. Selective attention has been studied with psychological methods for many years, but recent cognitive neuroscience studies using brain-imaging methods (and other neurobiological measures) have transformed the topic. Such studies have demonstrated that sensory processing can be strongly modulated by attention throughout perceptual networks, including even processing in primary sensory cortex. They have shown that some of these modulations can be anticipatory, arising prior to stimulus presentation, while other components reflect a changed neural response to an incoming stimulus. Recent imaging studies have also examined the mechanisms involved in controlling such attentional modulation of sensory processing. In addition to answering many long-standing questions about selective attention, such research also raises new questions. The various contributions in this volume provide an overview of the spectacular recent advances in attention research using neurobiological measures, and they outline critical issues to be resolved in the future.

Attention to action in Parkinson's disease: impaired effective connectivity among frontal cortical regions

The neural basis of attention to action was studied in 12 patients with Parkinson's disease (Hoehn and Yahr grades II and III) and 12 healthy age-matched controls. The subjects were studied by functional MRI (fMRI) during performance of a simple paced overlearned motor sequence task, with and without an additional attentional task. For the attentional task, subjects were instructed to attend either to their actions or to a visual distractor task. Statistical parametric mapping was used to implement a random effects analysis of the regional task-related activations in patient and control populations. Structural equation modelling of fMRI time series was used to measure effective connectivity among prefrontal and premotor areas. In both groups, the motor task was associated with activation of a distributed network including the premotor, motor and parietal cortex, striatum and cerebellum. In control subjects, but not patients, attention to action (relative to execution of an overlearned sequence) was associated with further activation of prefrontal, parietal and paracingulate cortex, and the supplementary motor area (SMA). Patients with Parkinson's disease showed greater than normal activation of the SMA during execution of the simple overlearned motor sequence, but less augmentation when attending to their actions. In control subjects, attention to action, but not attention to the visual distractor task, increased the effective connectivity between prefrontal cortex and both the lateral premotor cortex and the SMA. This represents a specific increase in effective connectivity. Attentional modulation of effective connectivity between the prefrontal, premotor cortex and SMA was not observed in patients. This deficit indicates a context-specific functional disconnection between the prefrontal cortex and the supplementary and premotor cortex in Parkinson's disease.

Simulating transcranial magnetic stimulation during PET with a large-scale neural network model of the prefrontal cortex and the visual system

Transcranial magnetic stimulation (TMS) exerts both excitatory and inhibitory effects on the stimulated neural tissue, although little is known about the neurobiological mechanisms by which it influences neuronal function. TMS has been used in conjunction with PET to examine interregional connectivity of human cerebral cortex. To help understand how TMS affects neuronal function, and how these effects are manifested during functional brain imaging, we simulated the effects of TMS on a large-scale neurobiologically realistic computational model consisting of multiple, interconnected regions that performs a visual delayed-match-to-sample task. The simulated electrical activities in each region of the model are similar to those found in single-cell monkey data, and the simulated integrated summed synaptic activities match regional cerebral blood flow (rCBF) data obtained in human PET studies. In the present simulations, the excitatory and inhibitory effects of TMS on both locally stimulated and distal sites were studied using simulated behavioral measures and simulated PET rCBF results. The application of TMS to either excitatory or inhibitory units of the model, or both, resulted in an increased number of errors in the task performed by the model. In experimental studies, both increases and decreases in rCBF following TMS have been observed. In the model, increasing TMS intensity caused an increase in rCBF when TMS exerted a predominantly excitatory effect, whereas decreased rCBF following TMS occurred if TMS exerted a predominantly inhibitory effect. We also found that regions both directly and indirectly connected to the stimulating site were affected by TMS.

Mesoscopic neurodynamics

An overview is given on the history, theory, and applications of mesoscopic neural models. It is shown why and how we can describe the behavior of a large population of neurons by the method of statistical neurodynamics. Normal and epileptic propagation of cortical activity have been simulated.

Functional interactions of the inferior frontal cortex during the processing of words and word-like stimuli

The hypothesis that ventral/anterior left inferior frontal gyrus (LIFG) subserves semantic processing and dorsal/posterior LIFG subserves phonological processing was tested by determining the pattern of functional connectivity of these regions with regions in left occipital and temporal cortex during the processing of words and word-like stimuli. In accordance with the hypothesis, we found strong functional connectivity between activity in ventral LIFG and activity in occipital and temporal cortex only for words, and strong functional connectivity between activity in dorsal LIFG and activity in occipital and temporal cortex for words, pseudowords, and letter strings, but not for false font strings. These results demonstrate a task-dependent functional fractionation of the LIFG in terms of its functional links with posterior brain areas.