Faculty Bibliography

When brain responses to sensory stimuli are averaged over time, non-timelocked neural activity is canceled out. Because sensory systems fluctuate in their reponsiveness over time, some of this non-timelocked ('induced') activity is likely to be functionally relevant for sensory processing. To analyze non-timelocked activity, we used multitaper spectral analysis and FFT to identify frequency components of single evoked responses. Frequency components of groups of single evoked responses were then analyzed by parametric and nonparametric statistical analyses. When applied to magnetoencephalographic (MEG) recordings, this method identifies which MEG sensors recorded statistically significant differences in specific frequency ranges in different simulus and response conditions. This method was applied to MEG recordings of human cortical magnetic responses evoked by pairs of brief sequential somatosensory stimulation of both index fingers. Intrapair intervals were 2-34ms (around the somatosensory flicker-fusion threshold). Subjects' button-press reports indicated whether tap pairs were perceived as single events (fused) or sequential events (flicker). MEG recordings of individual trials were grouped and analyzed according to (a) subjects' responses and (b) stimulus type. Preliminary results indicate significant differences (P<0.0005) in the amplitude of specific frequency ranges depending on whether the subjects reported that the tap pairs were perceived as flickering vs fused

Purkinje cell LTD was evoked in mouse cerebellar slices by pairing either a) direct b) glutamate iontophoresis, or c) climbing fiber activation with parallel fiber stimulation. Patch clamping was implemented at the soma which supported parallel fiber evoked synaptic current measurements as well as direct activation of the Purkinje cell. The stimulus paradigm consisted of low frequency pairing (1Hz) for a period of 5 minutes of the either of the three (a, b or c) activation modes with parallel fiber stimuli. Parallel fiber Purkinje cell LTD was evoked by all three paring paradigms and the depression was followed for periods of over one hour. Slice pre-incubation with the neuroprotection agent T-588 (Toyama Chemical Co., Ltd.), at concentration less than 1 microM, produced a total suppression of LTD under similar conditions as describe above. Two-photon microscopy determination of calcium concentration profiles in Purkinje cells demonstrated a good correlation between I Ca2+ increase and LTD. This calcium increase was prevented in the presence of T-588. The results suggest that LTD is triggered to prevent cellular damage due to excessive intracellular calcium concentration. Thus, the presence of T-588 obviates the necessity of triggering LTD neuroprotection. These results are particularly significant, as T-588 while preventing LTD in vitro, does not impair motor learning in pretreated animals

Thalamocortical dysrhythmias (TCD) may underlie a variety of neurological and neuropsychiatric symptoms. In TCD, pathological theta-range (4-8 Hz) activity from thalamic deafferentation or disfacilitation has been hypothesized to trigger thalamocortical (TC) domains to function at low frequency, surrounded by areas of gamma-band activity. This intersection has been viewed as creating an 'edge effect' which underlies some positive symptoms. TC properties could also maintain and distribute TCD. Spontaneous neuromagnetic activity was recorded from patients suffering from refractory obsessive-compulsive disorder (OCD) and/or major depression (MD) and from healthy controls. Recordings were performed with a whole-head magnetoencephalogram (MEG) (4-D Neuroimaging) in a shielded room. Activity was recorded for 5-10 min while subjects rested with eyes closed. Spectral analysis using a multitaper technique and cross-correlations between spectral amplitudes were calculated using Matlab (Mathworks, Inc.) and in-house software on a Linux cluster computer system. Power spectra from control recordings demonstrated typical alpha-rhythms, while spectra from OCD and MD subjects showed robust activity in the theta range and increased total power compared to controls. Coherence patterns from controls displayed activation of discrete frequency ranges, while patterns from OCD and MD subjects showed high coherence over a wide range of frequencies. This may reflect theta-range recursive corticothalamic activation

Historically different states of awareness have been correlated with coherent oscillatory neural activity at different frequencies. These brain rhythms are usually measured noninvasively with either EEG or MEG technology. However, few studies have attempted to localize the sources of the electromagnetic fields underlying these cognitive states. In this study, an anatomically constrained iterative source imaging method was used to characterize the current density distribution of the ganerators of the different brain rhythms. Magnetic activity was recorded with a 148-channel neuromagnetometer array (4D Neuroimaging) while subjects rested with eyes closed during wakefulness (WA), slow wave sleep (SWS), or REM sleep (RS). EKG and EOGs were recorded and used for artifact removal based on QRS modeling and ICA. Spectral analysis was performed using multitaper methods. The cross spectral density matrices, their SVD, and the coherence across frequencies were computed. Sources were analyzed for different frequencies. Spectral analysis indicates that during both WA and RS there is enhanced power in the alpha, beta, and gamma bands. During SWS there is enhanced delta, theta, and 14Hz sleep spindles, while beta and gamma are reduced. Source analysis indicates that alpha currents are distributed in occipital and parietal cortex during WA and RS, but shifts to the frontal lobe during SWS. Beta and gamma sources showed more complex distributions throughout different cortical regions. Sources of delta and theta activity were more widely distributed throughout the temporal, frontal, and occipital cortex. Slow delta was more frontal. Sleep spindles were localized to parietal cortex

Synchronous gamma-band neural activity has been proposed as a mechanism for 'binding' sensory stimuli into unitary conscious representations. Previous auditory and somatosensory experiments indicate covariance between gamma-band activity and subjects' thresholds for identifying one vs two sensory stimuli. Using a 148-channel whole-head MEG system (4-D Neuroimaging), we investigated whether similar covariance between perceptual thresholds and gamma-band activity is evident when human subjects identify stimuli that separately activate cortical hemispheres. Subjects were stimulated with taps of 2 ms duration from 2 piezoelectric stimulators, each held between the thumb and index finger of each hand. Both hands were stimulated synchronously with one tap each, or the left hand tap preceded the right hand tap by 2, 4, 6, 8, 10, 12, 14, 16, 18, 20, 22, or 24 ms. After each trial, subjects indicated whether the taps overlapped in time (perceived as one event) or were consecutive (perceived as two events). Evoked magnetic responses were averaged, analyzed by FFT, Principal Component Analysis, and filtered in several frequency ranges (1-400, 15-100, and 20-50 Hz). Results indicate a psychophysical threshold of apprx12 ms. For stimuli perceived as simultaneous, a single 20-50 Hz response occurred in each hemisphere. For taps perceived as sequential, a second gamma-band response was most evident in the hemisphere contralateral to the first stimulus. This suggests that somatosensory stimuli are bound across cortical hemispheres with a time-course similar to that of stimuli bound within a hemisphere, thus supporting the hypothesis that gamma-band activity is a neurophysiological correlate of perceptual binding