Hypotheses regarding the neural architecture that underlies human visual perceptions and object recognition have emerged from a large body of electrophysiological studies on non-human primates and other species. In particular it is proposed that visual information is hierarchically organized into stages that "build" more complex percepts. However, in humans direct investigations have not been possible until the development of fMRI. Guided by these hypotheses we investigated the cortical representations of illusory contours and stereo depth and determined that these "higher level" processes occur outside of primary visual cortex as predicted and also tend to be lateralized to the right hemisphere. Although it is not surprising that these percepts were represented beyond the primary levels, the right hemisphere assymetry was unexpected. A goal of future investigations is to understand the global organization including functional assymetries of visual specialization that leads to object recognition and vision-based cognition. As a foundation for these global questions, we have also identified frontal and parietal lobe systems involved in eye position, eye movements and perceptions of where things are located in space. Decision making and problem solving strategies that employ visual information will also be explored and related to the language studies. Each of these investigations also yields new options for functional localizations useful for neurosurgical planning where visual functions are at risk.

HUMAN LANGUAGE:
The discovery of language-specific areas.

Current models of language-related functional neuroanatomy are derived primarily from lesion data, direct cortical stimulation, and highly invasive preoperative and intra-operative procedures. However, apart from the involvement of specific cortical and subcortical structures of the language-dominant hemisphere, little is known about the cerebral mechanisms that mediate the acquisition, representation, and coordination of languages. Current studies in the fMRI lab at Columbia University address the question of how a second language is represented in the cortex relative to the first (native) language. Preliminary results suggest that the representation of a second language acquired in early adulthood is spatially separated from the native language in putative Broca's area; whereas, when acquired during the early language acquisition stage of development (presumed "sensitive period"), the two languages tend to overlap. Current and future basic science directions aim to further characterize the human language systems including object naming, language reception (auditory stimulation), verbal reasoning, and speech production. Neuroanatomical functional mapping of language-related cognitive activities provides another example of a function that has value for both the investigation of basic neuroscience questions and neurosurgical planning. We have developed an fMRI test to locate language-sensitive areas in the brain prior to surgery. Evaluations of the speech area localizations are made by direct comparisons with results of the WADA test (intracarotid amytal injection) to identify the dominant language hemisphere and post-surgical outcomes. Currently, all patients entered in the protocol show corresponding results between the conventional WADA test and the fMRI studies.

CHRONIC PAIN:
The discovery of areas sensitive to the relief of chronic pain.

Both the perception and the treatment of chronic pain involve complex neural and physiological systems. Unfortunately, little is known about human brain activity associated with chronic pain syndromes. However, the emergence of fMRI opens an unprecedented opportunity to test hypotheses about the cortical representations of various types of chronic pain, and to evaluate therapeutics for specific types of pain. Results of our initial investigations confirm that the cessation of chronic pain following a therapeutic treatment such as a nerve block is associated with the elimination of specific cortical activity found outside of the primary sensory areas and may target the specific systems involved in treatment of some types of chronic pain. Future aims are to identify specific cortical systems involved in the relief of pain.

UTILITY FOR NEUROSURGICAL PLANNING:
Application to patient care.

The Columbia University committment to translate basic scientific results into benefits for patient care whenever possible is exemplified in the fMRI research program where the above studies are unified under the goal of developing and evaluating a battery of standard tasks intended to mark the cortical areas associated with motor, sensory, language, visual, and auditory functions that might be at risk in a planned surgical procedure. The long term goal is to reduce the potential morbidity associated with surgical therapeutics. In this study all conventional mapping procedures (such as the WADA test and intraoperative cortical stimulation of sensory and motor areas) that are employed for a specific patient are used to evaluate a prediction based on the fMRI map. Pre and post-surgical outcomes are also compared with the fMRI prediction. Approximately 60 patients have been entered on this protocol and the preliminary summaries indicate a strong correspondence with conventional methods and outcomes. Based on the results to date, there is little doubt that, when completed, this comparison of conventional and fMRI brain mapping will document the utility of the methods for improved outcomes in neurosurgical procedures.