Barry Horwitz, Ph.D. - US grants

Our laboratory studies the relationship between what is observed in functional neuroimaging studies and the underlying neural dynamics. To do this, we had previously constructed a large-scale computer model of neuronal dynamics that performs a visual object-matching task similar to those designed for PET/fMRI studies. We extended the model so that it could also simulate auditory processing, thus allowing us to investigate the neural basis of auditory object processing in the cerebral cortex. This model relates neuronal dynamics of cortical processing of auditory spectrotemporal patterns to fMRI data. A review of both models can be found in Horwitz &Husain (2007). A number of hypotheses relating to the functional organization of the auditory object processing pathway were employed in constructing the auditory model. One of the most important of these hypothesized that neurons in the more anterior parts of the pathway should respond better to more complex auditory objects than to simple tones and sweeps. To test this hypothesis, we (Kikuchi et al., in press) recorded and analyzed auditory responses of single neurons in macaque monkeys from three different sectors distributed caudorostrally along the superior temporal plane (STP): Sector I, mainly area A1;Sector II, mainly area RT;and Sector III, principally the rostrotemporal polar area. Mean onset latency of excitation responses and stimulus selectivity to monkey calls and other sounds, both simple and complex, increased progressively from Sector I to III. Also, whereas cells in Sector I responded with significantly higher firing rates to the other sounds than to monkey calls, those in Sectors II and III responded at the same rate to both stimulus types. The pattern of results support the proposal that the STP contains a rostrally directed, hierarchically organized auditory processing stream, with gradually increasing stimulus selectivity, and that this stream extends from the primary auditory area to the temporal pole. The selective decrease in the responsivity of rostral STP neurons to other auditory stimuli implies that the majority of these cells, unlike those in area A1, are driven best by complex acoustic features rather than simple ones, in agreement with the hypothesis of our large-scale neural model (Kikuchi et al., in press). Another study focused on using fMRI to understand the neural basis of long-term audio-visual memory, which plays a central role in such tasks as naming and reading (Smith et al., 2010a). 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. Our laboratory also has performed studies to elucidate the neural basis of speech production and its disorders. The laryngeal motor cortex (LMC) is indispensable for the vocal motor control of speech and song production (for a review, see Simonyan and Horwitz, in press). Patients with bilateral lesions in this region are unable to speak and sing, although their nonverbal vocalizations, such as laughter and cry, are preserved. Despite the importance of the LMC in the control of voluntary voice production in humans, the literature describing its connections remains sparse. We used diffusion tensor probabilistic tractography and functional magnetic resonance imaging-based functional connectivity analysis to identify LMC networks controlling two tasks necessary for speech production: voluntary voice as repetition of two different syllables and voluntary breathing as controlled inspiration and expiration (Simonyan et al., 2009). Peaks of activation during all tasks were found in the bilateral ventral primary motor cortex in close proximity to each other. Functional networks of the LMC during voice production but not during controlled breathing showed significant left-hemispheric lateralization. However, structural networks of the LMC associated with both voluntary voice production and controlled breathing had bilateral hemispheric organization. Our findings indicate the presence of a common bilateral structural network of the LMC, upon which different functional networks are built to control various voluntary laryngeal tasks. Bilateral organization of functional LMC networks during controlled breathing supports its indispensable role in all types of laryngeal behaviors. Significant left-hemispheric lateralization of functional networks during simple but highly learned voice production suggests the readiness of the LMC network for production of a complex voluntary behavior, such as human speech. Our laboratory has also continued to develop new methods for employing brain fMRI data to evaluate how different brain regions interact with one another during the performance of sensory, motor and cognitive tasks (i.e., methods to calculate functional and effective connectivity). 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 developed 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 (Smith et al., 2010b). 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 used SLDS to model fMRI 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.