Faculty Bibliography

Voltage-sensitive dye imaging (VDI) of cortical activation patterns generated by electrical stimulation of thalamocortical afferents axons at different frequencies was studied, in vitro, using mouse brain slices. The study demonstrated that thalamocortical afferent axons stimulation could follow frequencies as high as 120 Hz without marked reduction. By contrast the power spectral density amplitude ratio in cortical layer 4 demonstrated a rapidly sigmoidal reduction at frequencies above 60 Hz. Similar findings were obtained with direct cortical afferent axons stimulation that obviated possible interactions at thalamic level. As pre-synaptic afferent field potentials, simultaneously recorded with the VDI at cortical level, followed higher stimulation frequency, it is concluded that cortical activity reduction is secondary to synaptic transmission failure. This interpretation agrees with the result from deep-brain stimulation (

Serotonin release from rat caudate preterminals and abdominal mast cell was determined in vitro using three-photon autofluorescence imaging. Three photon illumination was implemented with a Millennia X femtosecond laser < 80fs Tsunami using an excitation wave length of 750nm. Pulse broadening reduction was achieved via GDV (External Group Velocity Dispersion) compensation. An AX70 Olympus microscope with an x60 infrared objective lens, a homemade scanning/detection system, and data analysis program were a part of the three-photon system. Direct inspection demonstrated that serotonin (but not dopamine nor noradrenalin) filled micro-capillaries fluoresced at 750nm. Mast Cell: Fluorescent was observed in the large mast cell vesicles in the cytosol space. Following addition of 100uM ATP to the bath fluoresce was reduced bellow one third and the cell volume increased about 20%. No reduction of left over fluorescent was observed. Caudate Nucleus: bright fluorescent dots of about 1u diameter, which matched the size of pre-synaptic terminals surronded the somata of caudate neuron. Fluorescence intensity rapidly decreased after electric stimulation of adjacent area. Clear differences were detected in the fluorescence response to activation of mast cell and caudate nucleus tissue. In the former the number of fluorescent particles changed in the latter the overall intensity changed. This is to be expected given that the size of mast cell vesicle is sufficiently large to be imaged individually while this is not the case for presynaptic vesicles. In the latter case vesicular clusters are probably being imaged

Thalamocortical dysrhytmias (TCD) are at the basis of many pathological conditions that are characterized by the increment of low frequency oscillatory activity. Such increase could be generating asymmetrical neuronal activity within the boundaries of low and high frequency oscillating cortical neurons (i.e. edge effect; Llinas et al. 1999; PNAS). Here we studied the spatiotemporal interactions between high and low frequency inputs using optical imaging of voltage-sensitive dye signals combined with field potentials and intracellular recordings. Somatosensory cortical slices from 2-5 week old rats (350-400 mum thick) were stained with the voltage-sensitive dye RH795 and the fluorescent activity was recorded using a CCD camera (Fujix HRDeltaron 1700). Single or low frequency (2 to 10Hz) and high frequency (40Hz) repetitive stimulation were applied on cortical white matter.)Our results indicate that a single shock was able to generate a well-defined area of cortical activation with a radially oriented columnar configuration (n=8). On the other hand, the repetitive 40Hz stimulation was rapidly restraining the activation area to a smaller radial column (n=6) in agreement with previous results (Contreras & Llinas, 2001; J. Neurosci.). When a single shock was applied during the 40Hz stimulation, spatiotemporal profiles of the single shock cortical responses were unaffected. However, there was a single shock-dependent increment of activity within deep layers at the 40Hz cortical activation area (n=5). Thus, our results suggest that high-frequency cortical activation can be modulated by low frequency inputs

Individual evoked MEG or EEG responses typically exhibit low signal-to-noise ratios (SNRs) and high epoch variance across epochs; often hundreds of epochs are averaged to achieve adequate SNRs. High SNRs of gamma-band (20-50 Hz)activity were previously observed in the mean early (apprx50ms) evoked responses to brief tactile taps. However, SNRs are lower in evoked responses to pairs of stimuli presented near the subject's flicker-fusion threshold (apprx12 ms SOA) for pairs of identical tactile taps (or auditory clicks). At this threshold the gamma-band response to the first stimulus appears to be reset by the second stimulus, thus reducing SNRs. We attempted to increase SNRs in somatosensory evoked MEG responses to finger taps in 7 sighted and 7 blind subjects. Individual epochs were evaluated for frequency, phase, and amplitude similarities to the total average response during a window 20-90ms poststimulus. Epochs most similar to the total average were selected and averaged; the SNR of this response was compared to the SNR of the overall mean. ICA (Makeig, et al.) was also performed across all epochs and in conjunction with epoch selection. The two methods, applied separately or together, increased SNRs of the early evoked response. Detailed temporal structure of the evoked responses also became more evident, permitting increased resolution of the gamma-band reset near the subjects' psychophysical binding thresholds

Evoked magnetic fields were recorded while human subjects perceived apparent motion (AM) or 2 static bars of light separated by 1deg. Flash duration was 1ms and stimulus onset asynchrony (SOA) was 0 or 40ms. There were 3 tasks: left, right, or no AM. Detailed psychometric curves of the min SOA of AM were obtained separately. AM was perceived with a probability of .5 and 1 for SOAs of 15 and 40ms, respectively. Signals were processed with ICA for artifact removal, and then averaged. Amplitude and phase information in the time-frequency domain were estimated by the gabor wavelet transform. The mean sensor power revealed a stimulus-locked burst of gamma, beta and 10Hz activity. The AM conditions had enhanced and more long-lasting gamma power. Sparse source estimates were computed with an anatomically constrained recursively weighted minimum-norm algorithm modified to minimize the total current density. The source space was constrained to the subsampled cortical surfaces and to sparse subcortical dipoles. Source estimates were computed for each time point, each independent component, and for time-frequency subspaces of the wavelet transformed signals. Enhanced dynamic patterns in cortical areas V1, V3A, hMT+, LG, FG, LOS, IPS, STS, and other clusters in parietal, frontal, and temporal lobe, and in LGN, pulvinar, and superior colliculi correlated with AM. Dynamic gamma band sources were reconstructed in many of the motion-sensitive visual areas, particularly in V3A, hMT+, parietal and temporal cortex. Our results support a role for quantal gamma band and recurrent network activity in the creation of AM qualia by integrating bottom-up and top-down processes

Thalamocortical dysrhythmias (TCD) may underlie a variety of neurological and neuropsychiatric symptoms. (1,2) In TCD, pathological theta-range activity from thalamic deafferentation or disfacilitation is hypothesized to trigger thalamocortical (TC) domains to oscillate at low frequency, underlying negative symptoms, surrounded by areas of gamma-band activity, creating an 'edge effect' leading to some positive symptoms. TC connectivity and neuronal properties can distribute and sustain this pathological equilibrium. Spontaneous neuromagnetic activity was recorded from patients (n=5) with refractory OCD and from controls (n=4). Recordings were performed with whole-head MEG (4D Neuroimaging), for 5-10 min (0.1-508Hz) with subjects' eyes closed. Coherence, multitaper-based spectral and independent component analyses (ICA) were performed using Matlab (Mathworks) and in-house software. Power spectra from control recordings demonstrated typical alpha rhythms, while spectra from OCD subjects showed robust activity in the theta range and increased total power. In addition, cross-correlations of spectral amplitude from controls displayed activation of discrete frequencies; patterns from OCD subjects showed high coherence over a wider spectral range. Furthermore, ICA revealed components with theta-range spectral properties and dipolar positions consistent with aberrant resting cortical and basal ganglia oscillations. The conception of TCD may serve as a template for the study and treatment of neurological and psychiatric disorders. 1.Llinas,R et al (1999) PNAS 96:15222-7 2.Llinas,R et al (2001) Thal Rel Sys 1:237-44

We used somatosensory cortical slices from 2-5 week old rats to investigate the role of neuromodulators on the interactions between low (2 to 10Hz) and high (40 Hz) frequency cortical inputs. Slices were stained with the voltage-sensitive dye RH795 and their fluorescent activity was recorded using a CCD camera (Fujix HRDeltaron 1700). Electrical stimulation was delivered by two concentric bipolar electrodes positioned over the surface of white matter with a 1-2 mm separation.) Application of 5-20M of carbachol (a muscarinic ACh receptor agonist) facilitated both the low and high frequency-induced cortical responses. A significant increase in amplitude and duration of the responses within the deep cortical layers was observed (n=5). However, the characteristic spatial profiles generated by the low and high frequency cortical activation remained unmodified (see the accompanying abstract Urbano et al., 2002; Contreras & Llinas, 2001; J. Neuroscience). Indeed, during high frequency stimulation, the area of initial cortical excitation was laterally reduced to a small radial column. Furthermore, in the presence of carbachol a single shock applied during the 40Hz stimulation was able to further facilitate the cortical signals induced by the 40Hz train. A 50-75 % increase in the slope of the rising phase of the 40Hz train responses was observed after the single shock stimulation (n=6). These results suggest that interactions between high and low frequency cortical activity could be significantly influenced by the muscarinic cholinergic system

Pyramidal cells of layers 2 and 3 comprise one of the principal cell classes in the neocortex, but because of their small size the electrophysiological properties of their dendritic tree have only recently been accessible to direct measurement (Waters et al 2001). We have generated a detailed compartment model of layer 2/3 pyramidal cells based on recent data and used it to predict the response of these cells to synaptic inputs arriving at different regions of the dendritic tree. Distributions of dendritic Na+, K+ and Ca2+ conductances were constrained by the requirement that model electrophysiology fit the measured responses. For such distributions we found that simulated layer 1 synaptic input to the apical tuft was surprisingly ineffective in triggering somatic firing. This was because active propagation of excitation from the tuft inward to the soma rarely occurred in the model, consistent with our experimental findings. In contrast, the model showed that somatic firing was readily driven by input to the basal arbor. These predicted layer 2/3 pyramid input/output characteristics differ from those of layer 5 pyramidal cell models (Rhodes and Llinas 2001). In simulations, tuft input is more effective in layer 5 pyramids than layer 2/3 pyramids, whereas layer 2/3 pyramids are more responsive to feedforward synaptic input impinging upon the proximal arbor. We propose that feedforward and feedback streams of information in cortex may have complementary effects upon the microcircuitry of the column