Low power digital communication in implantable devices using volume conduction of biological tissues.

This work investigates the data communication problem of implantable devices using fundamental theories in communications. We utilize the volume conduction property of biological tissues to establish a digital communications link. Data obtained through animal experiments are used to analyze the time and frequency response of the volume conduction channel as well as to characterize the biological signals and noises present in the system. A low power bandwidth efficient channel-coded modulation scheme is proposed to conserve battery power and reduce the health risks associated.

Uncoupling and recoupling of autonomic regulation of the heart beat in pediatric septic shock.

Healthy physiological systems exhibit marked signal variability and complexity, whereas diseased systems generally show a loss of variability, decreased complexity ("decomplexification"), and increased regularity. The goal of this study was to evaluate the uncoupling and recoupling phenomenon in children with septic shock by observing serial changes in heart rate variability metrics. Data were collected from 7 children with septic shock by using the computer system in the Complex Systems Laboratory at Oregon Health Sciences University. Heart rate time series were constructed and analyzed by using the Hales Research System at intervals of 6 h during pediatric intensive care unit (PICU) hospitalization. These power spectral values were then plotted vs. time. Six of seven patients showed an increase over time in low-frequency heart rate power and the low-/high-frequency ratio, whereas high-frequency heart rate power decreased. We also compared the change in mean heart rate, heart rate standard deviation, and power spectral values during the first 24 h of PICU hospitalization vs. the remainder of the PICU stay (for the 5 patients with a PICU length of stay > 48 h). Compared to the initial 24 h in the PICU, low-frequency power and the low-/high-frequency ratio increased, whereas high-frequency power decreased over the course of the illness. This report shows the potential value of monitoring the uncoupling and recoupling phenomenon in patients with septic shock. Our results are in agreement with other investigators who report evidence of decomplexification both in experimental models of sepsis and in clinical studies and provide direction for further work.

EEG source localization: a neural network approach.

Functional activity in the brain is associated with the generation of currents and resultant voltages which may be observed on the scalp as the electroencephelogram. The current sources may be modeled as dipoles. The properties of the current dipole sources may be studied by solving either the forward or inverse problems. The forward problem utilizes a volume conductor model for the head, in which the potentials on the conductor surface are computed based on an assumed current dipole at an arbitrary location, orientation, and strength. In the inverse problem, on the other hand, a current dipole, or a group of dipoles, is identified based on the observed EEG. Both the forward and inverse problems are typically solved by numerical procedures, such as a boundary element method and an optimization algorithm. These approaches are highly time-consuming and unsuitable for the rapid evaluation of brain function. In this paper we present a different approach to these problems based on machine learning. We solve both problems using artificial neural networks which are trained off-line using back-propagation techniques to learn the complex source-potential relationships of head volume conduction. Once trained, these networks are able to generalize their knowledge to localize functional activity within the brain in a computationally efficient manner.

Real-time intraoperative neurophysiological monitoring.

We discuss the utilization of signal processing techniques during surgical procedures. These techniques are used to provide real-time monitoring of nervous system function. We describe the historical development of these techniques and the hardware and software that have been used to implement them.

Extraction and analysis of early ictal activity in subdural electroencephalogram.

Subdural electroencephalograms (SEEGs) are of great value in localizing primary epileptogenic regions in patients undergoing evaluation for focal resective epilepsy surgery. The data segments which contain a transition from interictal to ictal activity carry the most critical diagnostic information. Computer signal extraction within this transition period represents a challenging signal processing problem. In this work a two-step method is presented to extract early ictal activity. In the first step we employ a nonlinear signal decomposition technique in the wavelet domain to separate SEEG data into ictal and background components. In the second step we use time-frequency analysis and a novel integration algorithm to extract the desired information. Our experiments on clinically recorded data indicate that this method is highly effective allowing us to reveal important hidden features in the data which could not otherwise be observable.

Characterization of sleep spindles using higher order statistics and spectra.

This work characterizes the dynamics of sleep spindles, observed in electroencephalogram (EEG) recorded from humans during sleep, using both time and frequency domain methods which depend on higher order statistics and spectra. The time domain method combines the use of second- and third-order correlations to reveal information on the stationarity of periodic spindle rhythms to detect transitions between multiple activities. The frequency domain method, based on normalized spectrum and bispectrum, describes frequency interactions associated with nonlinearities occurring in the observed EEG.

The forward EEG solutions can be computed using artificial neural networks.

Study of electroencenphalography (EEG) is the one of the most utilized methods in both basic brain research and clinical diagnosis of neurological disorders. Recent technological advances in computer and electronic systems have allowed the EEG to be recorded from large electrode arrays. Modeling the brain waves using a head volume conductor model provides an effective method to localize functional generators within the brain. However, the forward solutions to this model, which represent theoretical potentials in response to current sources within the volume conductor, are difficult to compute because of time-consuming numerical procedures utilized in either the boundary element method (BEM) or the finite element method (FEM). This paper presents a novel computational approach using an artificial neural network (ANN) to map two vectors of forward solutions. These two vectors correspond to different head models but with respect to the same current source. The input vector to the ANN is based on the spherical head model, which can be computed efficiently but involves large errors. The output vector from the ANN is based on the spheroidal model, which is more precise, but difficult to compute directly using the traditional means. Our experiments indicate that this ANN approach provides a remarkable improvement over the BEM and FEM methods: 1) the mean-square error of computation was only approximately 0.3% compared to the exact solution; 2) the online computation was extremely efficient, requiring only 168 floating point operations per channel to compute the forward solution, and 10.2 K-bytes of storage to represent the entire ANN. Using this approach it is possible to perform real-time EEG modeling accurately on personal computers.

Comparison of orthogonal search and canonical variate analysis for the identification of neurobiological systems.

In this paper, we investigate two general methods of modeling and prediction which have been applied to neurobiological systems, the orthogonal search (OS) method and the canonical variate analysis (CVA) approach. In these methods, nonlinear autoregressive moving average with observed inputs (ARX) and state affine models are developed as one step predictors by minimizing the mean-squared-error. An unknown nonlinear time-invariant system is assumed to have the Markov property of finite order so that the one step predictors are finite dimensional. No special assumptions are made about model terms, model order or state dimensions. Three examples are presented. The first is a numerical example which demonstrates the differences between the two methods, while the last two examples are computer simulations for a bilinear system and the Lorenz attractor which can serve as a model for the EEG. These two methods produce comparable results in terms of minimizing the mean-square-error; however, the OS method produces an ARX model with fewer terms, while the CVA method produces a state model with fewer state dimensions.

Saccular aneurysm formation in curved and bifurcating arteries.

BACKGROUND AND PURPOSE: Distinguishing whether forces resulting from the impingement of central blood flow streams at a curved arterial segment or at the apex of an intracranial bifurcation could be important for the understanding of aneurysm formation. Using finite element models, our purpose was to investigate the hemodynamics related to intracranial saccular aneurysm formation through computer simulations. METHODS: We present two-dimensional finite element models describing several distinct stages of aneurysm formation in both curved and bifurcating arteries. For each model, a description of the numeric solutions and results are presented. RESULTS: Our results suggest that the pressures and shear stresses that develop along the outer (lateral) wall of a curved artery and at the apex of an arterial bifurcation create a hemodynamic state that promotes saccular aneurysm formation. The impingement of the central stream results in greatly increased velocity/pressure gradients and high shear stresses at the apex compared with those in the proximal parent or distal daughter branches. The results also indicate that the maximal pressure generated at the apex of the arterial bifurcation ranges from two to three times the peak luminal pressure in the proximal parent artery. CONCLUSION: These data suggest that, in the absence of any underlying disease process, aneurysm development is a mechanically mediated event. These models offer a plausible hypothesis regarding the initiation, growth, and subsequent rupture of saccular intracranial aneurysms as they relate to the hemodynamics of intracranial arterial blood flow.

Cervical spine evaluation in obtunded or comatose pediatric trauma patients: A pilot study.

A uniformly accepted protocol for evaluation and clearance of the cervical spine of pediatric trauma patients with altered mental status does not currently exist. We sought to detect cervical spine injuries in this group with minimal risk. Patients were evaluated with standard three-view cervical spine radiographs and CT when necessary. Those patients without radiographic abnormality and altered mental status underwent flexion-extension of the cervical spine using fluoroscopy with somatosensory evoked potential (SSEP) monitoring. Those with abnormal movement by fluoroscopy or changes in SSEP underwent MRI. Fifteen patients were evaluated with this protocol. Two patients had movement on flexion-extension of the cervical spine and 5 had SSEP changes. Three patients had an MRI with only 1 showing injury. Five patients had residual hemiparesis. Evaluation of the cervical spine in obtunded or comatose pediatric trauma patients can be done safely with flexion-extension under fluoroscopy and SSEP monitoring. Further prospective studies are required to determine the efficacy of SSEP monitoring for cervical spine clearance in this select population.

Cerebral monitoring by means of oximetry and somatosensory evoked potentials during carotid endarterectomy.

OBJECT: Cerebral ischemia that occurs during carotid endarterectomy is commonly monitored by means of somatosensory evoked potentials (SSEPs) and electroencephalography (EEG). The authors conducted this study to determine whether cerebral ischemia could also be reliably detected by cerebral oximetry. METHODS: Twenty-nine patients who underwent carotid endarterectomy were monitored by means of SSEPs, EEG, and cerebral oximetry with a model NIRO500 (20 patients) or INVOS3100A (nine patients) oximeter. Changes in amplitude of SSEPs were graded as follows: 0, no change; 1, decrease of less than 50%; 2, decrease of greater than 50%; and 3, 100% decrease. As measured with the NIRO500 oximeter, closing the common caro-tid artery decreased mean oxyhemoglobin levels twice as much (p < 0.005) in the group with SSEPs of 1 to 3 (-13.11+/-5.59 microM [mean+/-standard deviation], 12 patients) as in the group with SSEPs of 0 (-6.22+/-5.59 microM, eight patients). The rise in deoxyhemoglobin was also greater (p < 0.05). Two of nine patients monitored with the INVOS3100A oximeter had SSEPs of 1 and 3, and their regional saturation of oxygen (rSO2) values fell by -11.50 and -11.51, respectively. In the remaining seven patients with SSEPs of 0, the rSO2 ranged between -2.00 and -6.10 with no overlap with the group with SSEPs of I to 3. The increase in oxyhemoglobin monitored using the NIRO500 oximeter and rSO2 monitored using the INVOS3100A machine after opening the external carotid artery was less than that seen after opening the internal carotid artery. Both types of oximeters could detect cerebral ischemia but whereas false negatives occurred with the NIRO500, none was observed with the INVOS3100A. Extracranial contamination was also four times less frequent with the INVOS3100A than with the NIRO500 monitor. CONCLUSIONS: The results indicate that at least as measured with the INVOS3100A instrument, a decrease in rSO2 of -10 or more or a decrease below an rSO2 of 50 is indicative of cerebral ischemia of sufficient severity to decrease the amplitude of SSEPs.

Simultaneous somatosensory evoked potential and electromyographic recordings during lumbosacral decompression and instrumentation.

OBJECTIVE: Both motor and sensory neurological deficits have been reported after decompression and instrumentation of the lumbosacral spine. In this report, we describe a simple and effective method by which sensory and motor functions can be monitored simultaneously, using somatosensory evoked potentials (SSEP), spontaneous electromyographic (EMG) activity, and compound muscle action potential monitoring. The concomitant use of these monitoring techniques allows ongoing functional evaluation of the cauda equina and spinal cord during patient positioning, surgery, wound closure, and anesthetic emergence. METHODS: SSEPs were recorded continuously in response to peroneal or tibial nerve stimulation. EMG activity (both spontaneous and evoked) was recorded bilaterally from appropriate lower extremity muscle groups. All recordings (SSEP and EMG activity recordings) were obtained, stored, and reviewed simultaneously. RESULTS: SSEPs and EMG activity were simultaneously recorded for 44 patients. All patients in the study underwent surgical procedures to decompress and stabilize the lumbosacral spine, using pedicle screw instrumentation. In two cases, changes in SSEPs and spontaneous EMG activity were noted and were correlated with postoperative patient complaints. CONCLUSION: This report describes the concomitant use of powerful and simple tools that provide immediate, "early-warning" feedback to the surgical team concerning the sensory and motor functioning of the spinal cord and cauda equina. In addition, compound muscle action potential recording provides a tool for the identification of both levels and structures in the lumbosacral spine.

Real time signal processing in the clinical setting.

We routinely use a variety of real time signal acquisition, enhancement, and display techniques in the operating room to provide the surgeon with functional information. This enables reduction of surgical morbidity in cases which present a significant risk to the nervous system. Here we present regression based signal processing algorithms which produce considerable signal-to-noise-ratio enhancement with corresponding reduction in the time required to obtain an interpretable neurophysiological signal. We also present the approach we have applied to fault tolerance and distributed data display for our workstation cluster environment.

Effects of prenatal substance exposure: altered maturation of visual evoked potentials.

We investigated the effects of prenatal substance use on visual evoked potentials (VEPs). Seventy-four children were tested at birth and 1 month of age with binocular flash VEPs and at 4, 8, and 18 months of age with binocular pattern VEPs. Regressions were run by trimester to assess the independent effects of substance exposure. Variables included in the regression model were alcohol, marijuana, tobacco, other drug use for each trimester, maternal age, education, income, race, marital status, infant sex, birthweight, and Dubowitz score. Changes in specific components of the binocular VEP were both substance- and trimester-specific. First trimester alcohol use was associated with prolonged P1 wave latencies at 1 month of age. Prolonged P1 wave latencies at birth and 18 months were associated with tobacco use during each of the three trimesters, at 1 and 18 months with third trimester marijuana use, and at 1 and 18 months with first trimester other illicit drug use. Although these women were moderate substance users during pregnancy, their offspring exhibited maturational changes in components of the VEP in the absence of neonatal behavioral disturbances.

Regional differences in spectral EEG measures between healthy term and preterm infants.

State-specific spectral electroencephalographic (EEG) values were compared among 14 bipolar channel derivations between two healthy neonatal cohorts. Fifty-five healthy preterm neonates of < or = 32 weeks gestational age at birth were studied with 24-channel recordings over 3 hours at term conceptional age. These were compared with studies of 45 healthy term neonates. Five spectral measures for each channel (i.e., total spectral EEG, delta, theta, alpha, and beta frequency ranges) were calculated for each minute, which was identified as active or quiet sleep, based on visual analysis. Using multivariate analysis of variance, differences at each channel were assessed between neonatal cohorts for both states and cohorts; higher total EEG spectral values were noted during active sleep; whereas higher delta and theta spectral values were noted during quiet sleep. The term cohort had higher values for spectral theta, alpha, and beta power spectra in multiple channels, most significantly in the left central (i.e., C3O1) and sagittal regions (FzCz, CzPz) during both states (P < .0001, adj r2 > or = .2). Both interhemispheric and intrahemispheric differences in spectral values were present. For a healthy preterm cohort, lower spectral energies are expressed during sleep in specific head regions. Physiologic asymmetries exist in the newborn brain which are unique for the preterm infant, emphasizing functional alterations in brain development. How these asymmetries are altered by prenatal or postnatal stress or disease states needs to be explored.

Finite element methods in the simulation and analysis of intracranial blood flow.

This paper presents an introduction to the use of finite element methods in the simulation and analysis of intracranial blood flow and lays the foundation for more detailed clinically oriented studies. An overview of finite element theory is provided and includes the formulation of both the continuous and discrete equations of viscous fluid flow. A discussion of appropriate assumptions and boundary conditions governing arterial blood flow is presented. Two-dimensional, rigid-walled models are developed for flow in a straight artery, a 90 degrees curved artery and a bifurcated artery. For each model, a description of the finite element mesh, numerical solution and computational results are presented. This paper is the first in a series which will detail computational analysis of the relationship between pressure, velocity development of intracranial aneurysms and therapeutic approaches to aneurysm management. The goals of this research are to investigate the fluid dynamics that arise as a result of pulsatile flow in the arteries of the circle of Willis, relate these hemodynamics to the formation of aneurysms, develop a computational understanding of the effects of various therapies on blood flow related to aneurysms, and to develop and utilize patient specific computer simulations for treatment planning.

Computer classification of state in healthy preterm neonates.

Nineteen electroencephalographic (EEG) sleep measures describing four physiologic aspects of sleep behavior (i.e. sleep continuity, EEG spectra, body and eye movements, and autonomic measures) were derived from visual and computer analyses of 71 24-channel, 3-hour EEG sleep recordings on 52 healthy preterm neonates from 28-36.5 weeks postconceptional age (PCA). Forty-eight subjects were neurodevelopmentally normal up to 2 years of age. Four electrographic states that comprise tracé discontinu of the preterm neonate were defined in terms of increasing seconds of EEG quiescence per minute. A regression analysis was performed after transformations of nonlinear data sets representing the 19 EEG sleep measures, with the four sleep states as outcome variables. Postconceptional age was also included in these analyses as the 20th explanatory variable. Four measures best defined the EEG sleep states, explaining 75% of the variance: decreasing rapid eye movements per minute, decreasing numbers of spontaneous arousals per minute, increasing spectral theta energies, and decreasing facial movements per minute. Other cerebral and noncerebral measures, including total spectral EEG energies, spectral EEG energies in three bandwidths (i.e. delta, alpha, beta), cardiac and respiratory measures, and body movements, did not contribute as significantly to the prediction. Inclusion of PCA into the regression equation with the four EEG measures, selected by the analysis procedure, indicated that its contribution to state prediction was also small; the effect of PCA on state was found to be explained by the four EEG sleep measures.

Continuous intraoperative electromyographic monitoring of cranial nerves during resection of fourth ventricular tumors in children.

The authors reviewed the results of continuous intraoperative electromyographic (EMG) monitoring of muscles innervated by cranial nerves in 17 children whose preoperative imaging studies showed compression or infiltration of the fourth ventricular floor by tumor to determine how intraoperative EMG activity correlated with postoperative cranial nerve morbidity. Bilateral lateral rectus (sixth) and facial (seventh) nerve musculatures were monitored in all children. Cranial nerve function was documented immediately postoperatively and at 1 year. Of the 68 nerves monitored, nine new neuropathies occurred in six children (sixth nerve in four children and seventh nerve in five). In five new neuropathies, intraoperative EMG activity could be correlated in one of four sixth nerve injuries and four of five seventh nerve injuries. Electromyographic activity could not be correlated in four children with new neuropathies. Of 59 cranial nerves monitored that remained unchanged, 47 had no EMG activity. Twelve cranial nerves (three sixth nerves and nine seventh nerves) had EMG activity but no deficit. Of four children with lateral rectus EMG activity, three had new seventh nerve injuries. Lateral rectus EMG activity did not predict postoperative abducens injury. The absence of lateral rectus EMG activity did not assure preserved abducens function postoperatively. Likely because of the close apposition of the intrapontine facial nerve to the abducens nucleus, lateral rectus EMG activity was highly predictive of seventh nerve injury. Although facial muscle EMG activity was not an absolute predictor of postoperative facial nerve dysfunction, the presence of facial muscle EMG activity was associated statistically with postoperative facial paresis. The absence of facial muscle EMG activity was rarely associated with facial nerve injury. The authors speculate that EMG activity in the facial muscles may have provided important intraoperative information to the surgeon so as to avoid facial nerve injury.

Construction of 3-D arterial volume meshes from magnetic resonance angiography.

Finite element methods are well-suited for solving problems in arterial fluid dynamics, primarily due to their ability to handle flows in complex geometries. However, in order to use these computational methods to develop realistic models of pulsatile flow in intracranial arteries and associated aneurysms, it is necessary to construct a 3-D mesh, or grid, that accurately duplicates the arterial geometry of interest. In this paper, we present an efficient method to accurately develop realistic 3-D computational meshes of human intracranial arteries and aneurysms from serial magnetic resonance angiography images. However, these techniques may be applied to any other form of imaging data including computed tomographic angiography. First, raw grayscale images are segmented, converted to their binary form and arterial contours are extracted at each image slice. Next, the arterial contours are stacked and cubic splines are computed along the axial direction. This creates an affect similar to smooth integration in the axial direction and provides a set of points that define the 3-D arterial surface geometry. Then, surface patches are constructed and merged. A surface mesh is then computed with the ability to locally vary the mesh density as desired. Finally, nodal points on the surface mesh are used to compute the finite element volume mesh. The 3-D volume mesh accurately describes the arterial geometry and is used to develop patient-specific computational fluid dynamic models of flow phenomena in intracranial arteries and aneurysms. These flow models are then suitable for investigating the hemodynamics of intracranial aneurysm formation and test the end-effects of various medical and surgical treatments.

Continuous intraoperative electromyographic monitoring of cranial nerves during resection of fourth ventricular tumors in children.

The authors reviewed the results of continuous intraoperative electromyographic (EMG) monitoring of muscles innervated by cranial nerves in 17 children whose preoperative imaging studies showed compression or infiltration of the fourth ventricular floor by tumor to determine how intraoperative EMG activity correlated with postoperative cranial nerve morbidity. Bilateral lateral rectus (sixth) and facial (seventh) nerve musculature were monitored in all children. Cranial nerve function was documented immediately postoperatively and at 1 year. Of the 68 nerves monitored, nine new neuropathies occurred in six children (sixth nerve in four children and seventh nerve in five). In five new neuropathies, intraoperative EMG activity could be correlated in one of four sixth nerve injuries and four of five seventh nerve injuries. Electromyographic activity could not be correlated in four children with new neuropathies. Of 59 cranial nerves monitored that remained unchanged, 47 had no EMG activity. Twelve cranial nerves (three sixth nerves and nine seventh nerves) had EMG activity but no deficit. Of four children with lateral rectus EMG activity, three had new seventh nerve injuries. Lateral rectus EMG activity did not predict postoperative abducens injury. The absence of lateral rectus EMG activity did not assure preserved abducens function postoperatively. Likely because of the close apposition of the intrapontine facial nerve to the abducens nucleus, lateral rectus EMG activity was highly predictive of seventh nerve injury. Although facial muscle EMG activity was not an absolute predictor of postoperative facial nerve dysfunction, the presence of facial muscle EMG activity was associated statistically with postoperative facial paresis. The absence of facial muscle EMG activity was rarely associated with facial nerve injury. The authors speculate that EMG activity in the facial muscles may have provided important intraoperative information to the surgeon so as to avoid facial nerve injury.