Jitendran Muthuswamy

Higher-order spectral analysis of burst patterns in EEG

Burst suppression patterns in electroencephalograms (EEGs) have been observed in a variety of situations including recovery of a subject from a traumatic brain injury. They are associated with grave prognostic outcomes in neonates. The authors study power spectral parameters and bispectral parameters of the EEG at baseline, during early recovery from an asphyxic arrest (EEG burst patterns) and during late recovery after EEG evolves into a more continuous activity. The bicoherence indexes, which indicate the degree of phase coupling between two frequency components of a signal, are significantly higher within the /spl delta/-/spl theta/ band of the EEG bursts than in the baseline or late recovery waveforms. The bispectral parameters show a more detectable trend than the power spectral parameters. In the second part of the study, the authors looked into the possibility of higher (>2)-order nonlinearities in the EEG bursts using the diagonal slices of the polyspectrum. The diagonal elements of the polyspectrum reveal the presence of self-frequency and self-phase coupling of orders higher than two in majority of the EEG bursts studied. The bicoherence indexes and the diagonal elements of the polyspectrum strongly indicate the presence of nonlinearities of order two and in many cases higher, in the EEG generator during episodes of bursting. This indication of nonlinearity in EEG signals provides a novel quantitative measure of brains response to injury.

The use of fuzzy integrals and bispectral analysis of the electroencephalogram to predict movement under anesthesia

The objective of this study was to design and evaluate a methodology for estimating the depth of anesthesia in a canine model that integrates electroencephalogram (EEG)-derived autoregressive (AR) parameters, hemodynamic parameters, and the alveolar anesthetic concentration. Using a parametric approach, two separate AR models of order ten were derived for the EEG, one from the third-order cumulant sequence and the other from the autocorrelation lags of the EEG. Since the anesthetic dose versus depth of anesthesia curve is highly nonlinear, a neural network (NN) was chosen as the basic estimator and a multiple NN approach was conceived which took hemodynamic parameters, EEG derived parameters, and anesthetic concentration as input feature vectors. Since the estimation of the depth of anesthesia involves cognitive as well as statistical uncertainties, a fuzzy integral was used to integrate the individual estimates of the various networks and to arrive at the final estimate of the depth of anesthesia. Data from 11 experiments were used to train the NNs which were then tested on nine other experiments. The fuzzy integral of the individual NN estimates (when tested on 43 feature vectors from seven of the nine test experiments) classified 40 (93%) of them correctly, offering a substantial improvement over the individual NN estimates.

An array of microactuated microelectrodes for monitoring single-neuronal activity in rodents

Arrays of microelectrodes used for monitoring single- and multi-neuronal action potentials often fail to record from the same population of neurons over a period of time for several technical and biological reasons. We report here a novel Neural Probe chip with a 3-channel microactuated microelectrode array that will enable precise repositioning of the individual microelectrodes within the brain tissue after implantation. Thermal microactuators and associated microelectrodes in the Neural Probe chip are microfabricated using the Sandias Ultraplanar Multi-level MEMS Technology (SUMMiTV) process, a 5-layer polysilicon micromachining technology of the Sandia National labs, Albuquerque, NM. The Neural Probe chip enables precise bi-directional positioning of the microelectrodes in the brain with a step resolution in the order of 8.8 /spl mu/m. The thermal microactuators allow for a linear translation of the microelectrodes of up to 5 mm in either direction making it suitable for positioning microelectrodes in deep structures of a rodent brain. The overall translation in either direction was reduced to approximately 2 mm after insulation of the microelectrodes with epoxy for monitoring multi-unit activity. Single unit recordings were obtained from the somatosensory cortex of adult rats over a period of three days demonstrating the feasibility of this technology. Further optimization of the microelectrode insulation and chip packaging will be necessary before this technology can be validated in chronic experiments.

A study of electroencephalographic descriptors and end-tidal concentration in estimating depth of anesthesia

Objective. To study the usefulness of three electroencephalographic descriptors, the average median frequency, the average 90% spectral edge frequency, and a bispectral variable were used with the anesthetic concentrations in estimating the depth of anesthesia.Methods. Four channels of raw EEG data were collected from seven mongrel dogs in nine separate experiments under different levels of halothane anesthesia and nitrous oxide in oxygen. A tail clamp was used as the stimulus and the dog was labeled as a non-responder or responder based on its response. A bispectral variable of the EEG (just before a tail clamp) and the estimated MAC level of halothane and nitrous oxide combined were the two features used to characterize a single data point. A neural network analysis was done on 48 such data points. A second neural network analysis was done on 47 data points using average 90% spectral edge frequency and the estimated MAC level. The average median frequency of EEG was also evaluated, although a neural network analysis was not done.Results. The first neural network needed nine weights in order to train and correctly classify all of the 12 points in the training set under a training tolerance of 0.2. It could correctly classify all of the remaining 36 data points as either belonging to responders or non-responders. A cross-validation procedure, which estimated the overall performance of the network against future data points, showed that the network misclassified two out of the 48 data points. The second neural network needed 25 weights in order to train and classify correctly all of the 26 points in the training set under a tolerance of 0.2. It was later able to classify all of the 21 points of the test group correctly.Conclusions. The bispectral variable seems to reduce the nonlinearity in the boundary separating the class of non-responders from the class of responders. Consequently, the neural network based on the bispectral variable is less complex than the neural network that uses a power spectral variable as one of its inputs.

Ultrasound induced increase in excitability of single neurons

The aim of this study was to carefully assess the level of modulation in electrical excitability of single neurons with the application of high frequency ultrasound. High frequency tone bursts of ultrasound have been shown to dramatically increase the spike frequency of primary hippocampal neurons in culture. In addition, these ultrasonic bursts also induce silent or still developing neurons to fire. Results indicate that the increase in excitability is largely mediated by mechanical effects and not thermal effects of ultrasound. Future studies on culture models exposed to varying ultrasound protocols may provide insight into the feasibility of using ultrasound as a means for neurostimulation studies conducted on brain slice and in vivo models.

A chronic micropositioning system for neurophysiology

Microelectrode arrays fabricated for monitoring single and multi-neuronal action potentials often fail to record from the same population of neurons over a period of time. The number of recorded neurons tends to decrease due to micromotion of neurons away from the microelectrode, gliosis around the recording site and also brain movement due to behavior. We report here for the first time, a novel electrostatic microactuator based positioning system that will enable chronic tracking of single neurons. The electrostatic microactuators are fabricated using the SUMMiT/spl trade/ (Sandias Ultraplanar Multilevel MEMS Technology) process, a 5-layer polysilicon micromachining technology of the Sandia National Labs, NM. The microfabricated microactuators enable precise positioning of microelectrodes in the brain and subsequently upon implantation track specific neurons with minimal behavioral hindrance. The micropositioning system has been designed for chronic precision electrophysiology in rodents. We believe however, that this system will greatly enable precision monitoring and intervention during brain function and dysfunction.

Bispectrum analysis of EEG of a dog to determine the depth under halothane anesthesia

The solution of the problem of finding the depth of anesthesia or the level of unconsciousness by monitoring the electroencephalogram (EEG) is considered. The bispectrum analysis of the EEG of a dog is studied to determine any trends with increasing depth of anesthesia. Of the four channels of EEG analyzed, channel 1 (L/sub r/-L/sub 0/) seems to show a strong coupling between frequencies in the range of 6-10 Hz around 78% of the time when the dog is asleep (depth = 1.0) and a weak or no coupling between frequencies in the same range around 67% of the time when the dog is awake (depth = 0.0).<<ETX>>

Fabrication of screen-printed carbon electrode arrays for sensing neuronal messengers

Deciphering the methods of communication between neurons and ensembles of neurons in the brain is a major area of interest in the field of neuroscience. An array of sensors designed to sense specific neuronal messengers or neurotransmitters should provide a better method to study their spatial and temporal activity across a tissue. Screen-printing is a simple and inexpensive technique for fabricating arrays of sensors that can be used to monitor neurotransmitter activity in the brain. One important neuronal messenger known to actively modulate neuronal excitability is nitric oxide (NO). Carbon has been shown to interact with NO in an oxidation-reduction reaction that produces a current proportional to the amount of NO present. The proposed design uses carbon polymer inks screen printed onto aluminum traces to form the sensors. A thick, photodefineable epoxy resin, known as SU-8, serves as an insulator and a mold for the carbon ink. A potentiostat is used to apply a 900 mV voltage between the carbon sensor and a reference electrode positioned in the bath of the experimental setup. The current produced indicates the concentration of NO in close proximity to the carbon site. The screen-printing technique provides an elegant way to produce an array of individual carbon sensors. The carbon sensor array promises a novel approach to mapping the distribution of neurotransmitters in brain tissue.

Predicting depth of anesthesia using bispectral parameters in neural networks

Features like the spectral edge and median frequency derived from power spectrum of the EEG have so far failed to show any consistent changes with the depth of anesthesia. One of the disadvantages of using power spectrum is that it suppresses phase information in the signal. A third order spectrum or bispectrum preserves phase information. A bispectral parameter called bicoherence index was derived from the EEG prior to a tail clamp. Using the bicoherence index and the estimated MAC level of the dog at that time a neural network was able to correctly classify all the 36 data points from a test group corresponding to either an awake or an asleep dog.

Bispectral analysis of EEG burst patterns from piglets during recovery from asphyxia

The aim of this study was to study phase coupling among different frequency components in EEC burst patterns observed in piglets during recovery from asphyxia and to compare them with those of the awake baseline EEG data. Phase coupling was quantitated by determining the average bicoherence index within the delta band along the main diagonal in the bifrequency plane. Multiple instances of burst patterns from each experiment were analyzed for phase coupling and compared with the corresponding baseline EEG data. The baseline EEG data did not show any significant phase coupling within the delta band of frequencies in any of the experiments. The raw averaged bicoherence indices within the delta band of bursts was higher than that of corresponding baseline in almost all instances. The authors, therefore, conclude that the delta frequency components of the EEG tend to become more phase coupled during the initial phase of recovery of the brain from an asphyxic injury.