Natural visual scenes usually contain a large number of items and objects. Unfortunately, the processing capacity of our brain is inadequate to simultaneously parse all details of incoming information. This problem can be partially solved, however, by selecting the most relevant stimuli in a scene which will be processed in more detail compared to 'irrelevant' items. There is ample evidence that this process, called 'visuospatial attention', increases perceptual and neuronal sensitivity at the attended location. It has been suggested that 'top-down' signals originating in a fronto-parietal network of areas can modulate neuronal activity in occipital visual cortex via feedback connections. However, very few studies have actually shown a causal relationship between activity in this fronto-parietal network and modulations of activity in visual cortex. With support from the National Science Foundation, Dr. Wim Vanduffel and his group seek to investigate at a whole-brain scale the causal nature of functional interactions between frontal and occipital areas. Therefore, he will combine functional Magnetic Resonance Imaging (fMRI) techniques with electrical intracranial microstimulation (EM) in an animal model. Dr. Vanduffel will test whether the artificially increased output of a specific frontal area can lead to spatially specific alterations of sensory-driven activity in occipital cortex. If so, Dr. Vanduffel will investigate i) the extent of the functional network involved, ii) the stimulus specificity of the modulations, and iii) whether artificially induced modulations of sensory-driven activity mimic spatial attention effects. Finally, by combining fMRI with reversible deactivation techniques he will study whether specific parietal nodes within this functional network are crucially involved in the transfer of top-down information from frontal to occipital cortex.
The combination of the proposed experiments will move functional neuro-imaging from a mere mapping exercise to a real exploration of functional network dynamics, revealing mechanisms of cortical processing. Several undergraduate and graduate students will be trained during the activities sponsored by the National Science Foundation. Also, the proposed research will have a broader significance for fundamental research as the novel EM-fMRI method will provide a tool for detailed investigation of the functional role of feedback connections. These connections are poorly understood though increasingly viewed as crucial for understanding functional brain properties. The experiments may also have considerable clinical relevance for neurosurgeons performing peri-operative electrical stimulations. Moreover, the results of this project will also extend our theoretical understanding of attention deficits and may lead to the development of new analytical and diagnostic tools for attentional disorders.
We aimed to investigate, using causal rather than correlative methods, if and how signals originating in frontal cortex (in particular the frontal eye fields, FEF) are able to modulate activity throughout occipital visual cortex. We also investigated whether a region in parietal cortex (LIP) may be crucial in gating these presumptive feedback signals from frontal cortex towards visual cortex. In order to achieve these goals, we developed novel combinations of functional imaging and reversible perturbation methods. We mainly relied on functional magnetic resonance imaging (fMRI) using animal models, which allowed us to obtain brain-wide indirect measures of neuronal activity (through a hemodynamic signal). In addition, we combined fMRI with focal and reversible perturbation tools such as electrical microstimulation, and chemical inactivation. These tools allowed us to artificially increase or decrease neuronal activity of small cortical regions and investigate whether a region that is perturbed is crucially involved during task performance. Moreover, since we simultaneously measured fMRI activity we could correlate perturbation-induced changes in behavior with changes in brain activity. First, we improved the quality of our imaging tools by developing dedicated antennas (transmit-receive coils) needed to acquire the fMRI data. This included the development of external and implanted phased array coils which improved the spatial resolution of the measurements by more than a log unit (i.e. from 8 mm3 to sub-mm even to 0.125 mm3 voxels). The novel imaging technology that emerged from this project has been used in other studies which yielded already more than 40 publications. For example, we obtained very high resolution topographic maps of visual cortex (i.e. how the visual field is mapped in the cortex). Such detailed retinotopic maps are critical to identify those areas that are affected by artificially perturbed activity in frontal cortex. Next, we showed that artificially increased output of the FEF modulates fMRI activity throughout occipital cortex in a surprisingly heterogeneously manner: both regions showing increased and decreased fMRI activity levels were observed. Gamma-band activity and single unit activity within parietal area LIP showed a spatially- and task- specific modulation of activity. Moreover, we showed that the earliest stages of visual cortex were only affected by FEF stimulation in the presence of a retinal stimulus, suggesting that bottom-up and top-down signals interact. Furthermore, the observed fMRI effects in visual cortex (after FEF stimulation) were also stimulus-type dependent. The processing of low but not high salient stimuli improved. When comparing the effects of FEF microstimulation between a passive viewing task and a visually-guided saccade task we observed that FEF-induced modulation of fMRI activity in occipital cortex is not only stimulus but also task dependent. Hence coincident bottom-up and top-down activity increases the modulatory effects of FEF microstimulation on occipital cortical activity. In the next aim we tested the hypothesis that FEF-EM induced modulation of visually driven activity in occipital cortex is qualitatively similar to attention-dependent modulations of activity in these areas. As already mentioned above, we showed that microstimulation of the FEF had a very similar effect on contrast-varying stimuli positioned in the stimulated FEF-movement fields -as also reported in several selective spatial attention studies. We observed the strongest positive effects on stimuli with low luminance contrast but no (or even a suppressing) effect for high contrast stimuli. Furthermore, we scanned two animals performing a luminance contrast detection task, as used in previous FEF stimulation experiments (Moore and Fallah, 2004). After stimulation of the FEF, we observed both increased and decreased contrast detection thresholds when the target is presented in the stimulated FEF movement field. The effect depended on the location of the stimulation electrode (an electrode evoked either increased or decreased thresholds) and not the precise alignment of the stimulus relative to the stimulated movement field. This surprising result may suggest that different populations of FEF neurons influence occipital cortex in a different manner. This finding is also in line with the aforementioned data showing both enhanced and decreased fMRI activity in occipital cortex after stimulating the FEF. In the final aim we tested subjects that performed four cognitive tasks (variants of search and detection tasks) while (LIP) was reversibly inactivated and brain-wide fMRI activity was measured. Profound muscimol-induced increases in task-related fMRI activity were observed both in up- and downstream areas relative to LIP. These changes in fMRI activity were task-dependent and correlated with changes in behavioral performance. These results indicate that rapid adaptive mechanisms are triggered in areas remote from the inactivation site.
