Brain network interactions in auditory, visual and linguistic processing
Introduction
The study of human cognition, especially those cognitive functions such as language that are not found in other species, has been revolutionized by the introduction of functional brain imaging (fMRI, PET, EEG/MEG).1 One important feature of these types of data is that they are obtained essentially simultaneously from much of the brain and thus enable one to investigate not just what a single brain area does, but also how brain regions work together during the performance of individual cognitive tasks. For this reason, they are unlike the more traditional methods used to understand the neural basis of cognition (e.g., lesion analysis, single unit recording in nonhuman primates), which investigate one “object” at a time (e.g., the ideal brain damaged patient has a single localized brain lesion; single unit recordings are obtained from individual neurons in one brain region).
As a result, the complex data produced by functional brain imaging generates the need for network analysis and interpretation. Our research groups have undertaken the investigation of network interactions by combining computational neuroscience techniques with functional neuroimaging data. These analysis methods may allow us to evaluate how brain operations, in terms of network behavior, differ between tasks, and differ between normal and patient populations. Because of the latter, they may permit us to determine which networks are dysfunctional and the role neural plasticity plays in enabling compensatory behavior to occur. Central to this research is the use of large-scale biologically realistic network models that relate neuroanatomical and neurophysiological data to the signals measured by functional brain imaging. Not only does computational modeling help interpret the meaning of functional brain imaging data, it also provides a framework to generate and quantitatively test hypotheses concerning how specific cognitive tasks are implemented in the brain.
In this paper, we will illustrate our approaches to network analysis by reviewing several PET and fMRI studies that employed language and auditory tasks. In the next section we shall illustrate how functional neuroimaging data can be used to estimate how different brain regions are working together. Specifically, we use PET data to examine the functional connections between anterior and posterior perisylvian areas during language and language-related production tasks. In the following section, we illustrate the large-scale neural modeling approach to examine visual and auditory object processing.
Section snippets
Functional connectivity and language production
Although there appear to be a multitude of methods to analyze PET and fMRI data, almost all can be placed into one of two groups, each group being grounded on a different fundamental assumption concerning what brain regions do (Horwitz, 1994). The first assumption, functional specialization, proposes that different brain regions are engaged in different functions, and is implemented by comparing the functional signals between two scans (in its most simple formulation), each representing a
Network analysis – determining the neural substrate of specific cognitive tasks
In the last section, we presented an analysis of language production data that included discussions both of PET activations and of interregional functional connections. Even though we talked about areas where rCBF was greater in one condition than another, or areas where rCBF was related to rCBF in other areas, we discussed these results in terms of neural functioning and in terms of neural networks. However, relating PET or fMRI hemodynamically-based measurements to neural activity is
Conclusions
In this paper, we focused on the centrality of network interactions between brain regions for mediating the performance of sensorimotor and cognitive tasks. We emphasized functional neuroimaging, especially PET and fMRI, because these techniques provide data that are obtained simultaneously from much of the brain, thus allowing one to assess interregional functional interactions. In the first part of this paper, we demonstrated that frontal and posterior superior temporal gyral perisylvian
Acknowledgments
We wish to thank members of our laboratories for their efforts in the studies reported here. In particular, we thank Drs. Malle Tagemets and Fatima Husain for their work on the simulations we presented, Keith Jeffries for aid with the covariance analysis, and Dr. Fatima Husain for leading the experimental work associated with the auditory model. We are grateful to Dr. Carol Frattali for carefully reading the manuscript.
References (31)
- et al.
Auditory speech processing in the left temporal lobe: an electrical interference study
Brain and Language
(1995) The elusive concept of brain connectivity
NeuroImage
(2003)- et al.
Functional connectivity among language areas during speech production
NeuroImage
(2000) - et al.
Neural modeling, functional brain imaging, and cognition
Trends in Cognitive Sciences
(1999) - et al.
Does measurement of regional cerebral blood flow reflect synaptic activity? – Implications for PET and fMRI
NeuroImage
(1995) - et al.
Auditory processing in primate cerebral cortex
Current Opinion in Neurobiology
(1999) - et al.
Synthetic PET: analyzing large-scale properties of neural networks
Human Brain Mapping
(1995) - et al.
The neural organization of discourse: an H215O-PET study of narrative production in English and American sign language
Brain
(2001) - et al.
Altered patterns of cerebral activity during speech and language production in developmental stuttering. An H215O positron emission tomography study
Brain
(1997) - et al.
Large-scale neural model for visual attention: Integration of experimental single-cell and fMRI data
Cerebral Cortex
(2002)