Patent No. 6873872 Adaptive electric field modulation of neural systems

 

Patent No. 6873872

Adaptive electric field modulation of neural systems (Gluckman, et al., Mar 29, 2005)

Abstract

The present invention relates to devices and methods of modifying the neuronal activity of a neural system comprising neurons, comprising, one or more of the following steps, measuring the neuronal activity of a neural system; and applying an oriented electric field to said neural system effective to modify the neuronal activity of the neural system, wherein the magnitude and polarity of said applied electric field is changed in response to the measured neuronal activity. The present invention also relates to devices and methods for treating brain disorders, such as epilepsy and Parkinson's disease, comprising, one or more of the following steps, applying a sub-threshold and oriented electric field in situ to the brain of a patient having such a disorder in an amount effective to reduce the abnormal activity of the brain, wherein the electric field is applied through field electrodes in contact with the brain. The present invention also relates to methods and devices for restoring or repairing a brain function, such as sensation (e.g., taste, or smell), somatic activity, auditory activity, visual activity, or motor activity. It can also be used for testing drugs, pharmacological agents, and other modulators of neuronal function.

Notes:

 

GOVERNMENT INTERESTS STATEMENT REGARDING GOVERNMENT RIGHTS

This invention was made with government support under NIH Grant Nos. R01MH50006 and K02MH01493 awarded by the National Institution of Health. The government has certain rights in the invention.

Parent Case Text

This application is a continuation-in-part of U.S. Ser. No. 09/729,929, filed Dec. 6, 2000 now U.S. Pat. No. 6,665,562, which claims the benefit of provisional application Ser. No. 60/169,280, filed Dec. 7, 1999, which are hereby incorporated by reference in their entirety.

BACKGROUND OF THE INVENTION

Numerous attempts have been made to suppress epileptic seizures in human patients with indirect electrical stimulation at sites remote from the epileptic focus, including cerebellum (Cooper et al., 1976; Van Buren et al., 1978), thalamus (Cooper et al., 1985; Fisher et al., 1992), and vagal nerve (Murphy et al., 1995; McLachlin, 1997). Surprisingly, there has been far less investigation of the technology required to directly control an epileptic focus electrically. It has been shown that direct current injection into tissue could suppress evoked (Kayyali and Durand, 1991) or spontaneous (Nakagawa and Durand, 1991; Warren and Durand, 1998) epileptiform activity in brain slices. Even simple periodic pacing of a neuronal network with direct electrical stimulation (Kerger and Schiff, 1995) can reduce seizure-like events. In addition, there is some evidence that nonlinear control schemes might be useful in manipulating epileptiform activity (Schiff et al., 1994). In each of these cases, the stimulation was applied in the form of short current pulses directly into the tissue that evoke neuronal firing. Recently, it was demonstrated that steady state (DC) electric fields oriented parallel to pyramidal cells were capable of suppressing epileptic seizure activity in in vitro hippocampal brain slices (Gluckman et al., 1996a). Such field application led to nearly complete suppression of neuronal activity, yet due to a combination of polarization effects (electrode and tissue) and neuronal adaptation, this effect was transient.

DESCRIPTION OF INVENTION

The present invention relates to devices and methods for modulating the neuronal activity of a neural system comprising neurons, such as a brain, brain regions, or any in vivo or in vitro collection of neurons. In particular, the present invention involves the use of applied electric fields to modulate the behavior of a target neural system. In preferred embodiments, the polarity and magnitude of the applied electric field is varied according to information gathered from the modulated neural system, or any other desired source chosen to provide feedback, to modulate the strength of the applied electric field. In such embodiments, preferably a sub-threshold stimulus is administered to modulate to the neural system. The methods and devices of the present invention can be used to treat diseases of the nervous system, to restore neuronal function, paralysis, and motor and sensory deficits, to produce prosthetic devices that interact and modulate neuronal activity, to enhance or suppress neuronal activity and associated phenotypes, and the like.

A preferred method of the present invention relates to modifying the neuronal activity of a neural system comprising neurons, comprising one of more of the following steps, in any order: measuring the neuronal activity, or other behavior, of a neural system; and applying an oriented electric field to said neural system effective to modify the neuronal activity of the neural system, wherein the magnitude and polarity of said applied electric field is changed in response to the measured neuronal activity.

A neural system in accordance with the present invention can be any ensemble of one or more neurons, and/or other excitable cells, such as muscle, heart, retinal, cochlear, tissue culture cells, stem or progenitor cells, including cell-electrode interface devices and the like. Cells can be coupled electrically, chemically, or combinations thereof. The neural system can be an entire brain, ganglia, nerve, etc., or it can be a region or portion of it. Any animal source of material is suitable, including neural systems of invertebrates, such as mollusks, arthropods, insects, etc., vertebrates, such as mammals, humans, non-human mammals, great apes, monkeys, chimpanzees, dogs, cats, rats, mice, etc. In the examples, a specific region of a mammalian brain is dissected out and placed in a chamber where its activity is modified. However, physical isolation of a target brain region is unnecessary; the activity modulation can be performed in situ, as well. Preferred target regions include, but are not limited to, neocortex, sensory cortex, motor cortex, frontal lobe, parietal lobe, occipital lobe, temporal lobe, thalamus, hypothalamus, limbic system, amygdala, septum, hippocampus, fornix, cerebellum, brain stem, medulla, pons, basal ganglia, globus pallidum, striatum, spinal cord, ganglion, cranial nerves, peripheral nerves, retina, cochlea, etc.

In one step of a preferred method, the neuronal activity of the neural system is measured. By the term "neuronal activity," it is meant any measurable physical behavior, output, or phenotype of the system. For example, neurons typically display variations in their membrane potential, such as action potentials, depolarizations, and hyperpolarizations. These changes in the membrane potential can be utilized as a measure of neuronal activity, e.g., by monitoring intracellularly in a single neuron, or extracellularly, the electrical activity of a single neuron or the activity of an ensemble of neurons. Behaviors, or other products of a neural system (e.g., hormones, growth factors, neurotransmitters, ions, etc.) can also be detected, and used as a feedback signal to determine the magnitude and strength of the modulating applied field. For instance, if a purpose is to elicit movement of a limb, then the neuronal activity can be limb motion. The neuronal activity which is measured or assessed can be a subset of the total activity observed in the system, e.g., a particular frequency band of the full neural signal. In the examples, hippocampus slices were monitored for neuronal activity. Although the measuring electrode detected various types of activity, including spontaneous neuronal firing, slow burst activity, and background noise, as well as fast frequency epileptic seizures, it was desired to modulate only the latter. Thus, for these purposes the neuronal activity can be considered to be only the events of interest, e.g., the epileptic seizures.

Methods for measuring and recording neuronal activity can be accomplished according to any suitable method. In preferred embodiments of the invention, the neuronal activity is monitored extracellularly by measuring the extracellular electrical potential of a target population of neurons. Such measurements can reveal complex spikes or burst activity, sharp or slow waves, epileptiform spikes or seizures, arising from one or more neurons in the neural system.

The neuronal activity can be measured by recording the neural system's electrical potential in the extracellular space. The electrodes used to measure the field potential produced by the neural system are referred to as "measuring electrodes" or "recording electrodes." One or more electrodes can be used to measure the field potential. In preferred embodiments, two or more electrodes are utilized. The field potentials recorded at a given extracellular site will depend on a variety of factors, including the location of the electrode(s) with respect to the soma and dendritic layers, the architecture of the neural system, the perfusion solution, etc.

The measuring electrodes can detect the field potential from the applied field as well as the activity generated by the neural system. There are a number of methods that can be used to distinguish the neuronal activity from the applied fields. For example, in in vitro hippocampal slices, a pair of differential electrodes, aligned as closely as possible to the isopotential of the applied field, were used as measuring electrodes. They are "differential" in the sense that an active electrode is placed in the tissue, preferably near the cell body layer of the target neurons, while the reference electrode is placed preferably in the bath external to the tissue. The values obtained from each electrode can be electronically subtracted from each other, reducing background noise. For in vivo use, the differential measuring electrodes can be placed at the same isopotential with respect to the applied field. The electrodes can be as close to the target population as possible, without damaging it. Other methods to reduce noise and the artifact from the applied field can be used as well, either alone, or in combination with the differential electrodes, including filtering and post-processing of the measured signal.

The signal recorded from the system can be processed to dissociate the applied field potential from the electrical activity expressed by the neurons. Especially when the neuronal activity is to be used as a feedback signal to adaptively modulate the system, eliminating the electrical noise and artifact generated by the applied field may be an important factor in determining how to change it in response to the system's activity.

As mentioned above, placing the electrodes on an isopotential is one way to remove the applied field potential from the recorded signal. However, constraints imposed by the geometry and accessibility of the neural system may make placement of the electrodes along the isopotential impractical. In those cases, signal processing can be utilized. In simple terms, the goal of signal processing is to cancel any electrical field potential component, and associated noise, generated by the applied field input, so that the measured output can be attributed substantially to the activity of the neurons. Any processing method or technology which accomplishes this task can be used. The invention is not limited to how the processing is actually implemented, or the approach that is used.

The field potential contributed by the applied field can be determined using a relatively simple method. It is well-established that the potential difference between two points in a resistive medium generated by an applied field is proportional to the applied current used to generate the field. The proportionality constant will depend upon the specific properties of the system, e.g., its geometry, the conductive characteristics of the cells and tissues, the amount of fluid present, etc. Once the proportionality constant is known, the amount of potential produced by the field can be calculated, and then subtracted from the recorded signal. The proportionality constant can be determined mathematically from computer models accounting for the properties of the neural system, or it can be derived empirically. To make an empirical determination, a test signal can be inputted into the system, and then the resulting output can be measured. These two values (i.e., test signal and associated output) can be used to determine the constant. The derived proportionality constant can be used in the form of an algorithm to process the recorded signal. The algorithm can be implemented in any suitable hardware or software form.

Recording from the electrodes can be performed routinely. For instance, measurements can be made with an AC amplifier if the frequency and number of extracellular bursts are of interest. It can be equipped with filters to cut off frequencies below and above a particular range (band-pass filter) and amplify the signal in preferred ranges, e.g., 50-1000 Hz, preferably, 100-500 Hz. A DC amplifier can also be used, if slower potential changes are of interest.

A method in accordance with the present invention also involves applying an oriented electric field to the neural system effective to modify the neuronal activity of the neural system, preferably where the magnitude and polarity of said applied electric field is changed in response to the measured neuronal activity. Preferably, the applied field is oriented in a particular direction with respect to the somatic-dendritic axis of the neurons in the neural system. Most preferably, the field is parallel to the somatic-dendritic axis. Changing the strength of the applied field in response to a measured activity of the neural system can also be referred to as "adaptive modulation" since the strength of the applied field is adjusted based on an activity value of the neural system (e.g., electrical activity, motor activity, such as limb motion, etc.). A function of the applied electric field is to modify the neuronal activity of the neural system. The electric field is thus applied to the neural system in an amount adequate to change the neuronal behavior of the neural system. Any amount of field which changes the neural system's behavior is an effective applied field. It is believed that a mechanism that underlies adaptive modulation is the ability of the applied field to alter the neuron's excitability by changing its threshold; however, the invention is not bound nor limited to any theory, explanation, or mechanism of how it works.

In preferred methods of the present invention for in vitro applications, two pairs of electrodes can be used in the field application step. A pair of "field electrodes" can be used to produce the applied field. A second pair of electrodes, "sensing electrodes," can be used to measure or sense the field generated by the "field electrodes." The sensing and field electrodes can comprise the same materials described above for the measuring electrodes. In certain applications, however, such as in vivo applications, a field can be applied without sensing electrodes.

In preferred embodiments of the invention, the effective amount of applied field is sub-threshold with respect to the field potential experienced by the neural system. By the term "sub-threshold," it is meant that the amount of applied field or current does not reliably, with 100% probability, initiate new action potentials within the neural system. In contrast, the application of a supra-threshold stimulus reliably, with a high degree of probability, results in neuronal firing. A sub-threshold potential is, for example, less than 100 mV/mm, preferably 50 mV/mm and less, more preferably, 25 mV/mm and less, such as 20 mV/mm, 15 mV/mm, or 10 mV/mm. The sub-threshold potential refers to the potential generated at the level of the target neurons. The amount of potential actually produced by the field electrodes is less important that the field perceived by the target neurons. It is the generated field sensed by the neurons that determines whether a stimulus is sub- or supra-threshold.

In response to the applied electric field, the activity of the neural system can be modified in any desired manner, e.g., the activity can be suppressed, reduced, decreased, diminished, eliminated, counteracted, canceled out, etc., or it can be enhanced, increased, augmented, facilitated, etc. To determine whether the activity of the system has been modified, preferably the same neuronal activity measured in the measurement step is re-measured. Most preferably, the measurement of the neuronal activity is performed simultaneously and continuously with the applied field.

Any effective electrodes can be used for the recording, sensing, and field electrodes, including, e.g., metal, steel, activated iridium, platinum, platinum-iridium, iridium oxide, titanium oxide, silver chloride, gold chloride, etc., where the electrode can be insulated by glass or lacquer, as well as silicon microelectronics, including tetrode or other multielectrode arrays or bundles, multichannel and ribbon devices. Typically, the electrodes can have relatively large tips with low resistance to detect activity from a number of neuronal elements within the neural system. Smaller tipped electrodes can be used for monitoring activity from single neurons or smaller populations. Activity can be measured from one or more electrodes, preferably two or more. In some cases, it may be desired to record from several regions of the neural system in order to characterize its activity. Recordings of intracellular, extracellular, or a combination thereof, can be analyzed separately, or together. The electrodes can be AC- or DC-coupled.

For certain purposes, iridium oxide type electrodes may be preferred since they are relatively nontoxic to cells, as well as being effective carriers of high current and charge densities. An activated iridium or iridium alloy wire can be used, or a metal substrate, such as noble metal (e.g., Au, Pt, or PtIr), ferrous steel alloy, stainless steel, tungsten, titanium, Si microprobe, etc., or other suitable substrate, can be coated with a film of iridium oxide to produce an effective electrode. Any suitable method to prepare the coating can be used, including, but not limited to, an activation process (e.g., Loeb et al., J. Neuro. Sci. Methods, 63:175-183, 1995; Anderson et al., IEEE Trans. Biomed. Eng., 36:693-704, 1989) to form activated iridium oxide films (AIROFs), thermal decomposition (Robblea et al., Mat. Res. Soc. Symp. Proc., 55:303-310, 1986) to form thermal iridium oxide films (TIROFs), reactive sputtering (15) to form sputtered iridium oxide films (SIROFs), electrodepositing (Kreider et al., Sensors and Actuators, B28:167-172, 1995) to form electrodeposited iridium oxide films (EIROFs), etc.

As described herein, it has been found that adaptive modulation of a neural system can be used to modify its neuronal activity. In preferred embodiments, this is achieved by characterizing the neuronal activity and then using a feedback algorithm to determine the field magnitude necessary to modulate its activity. Neuronal activity can be characterized by various measurements, depending upon the particular activity that is being assessed. When electrical activity is a determinant, then measurements can include, e.g., local field polarity and magnitude (e.g., -10 mV), burst activity, burst amplitude, burst frequency, power in a predetermined frequency band of activity, non-burst activity, single or small population firing rate, amplitude or phase of periodic activity, such as theta rhythm, root-mean-square (RMS), variance, etc. In general, any suitable measure of neuronal activity can be used as the feedback stimulus for the applied field. The feedback stimulus can also be determined by multiple measurements, e.g., electrical activity, limb motion, cochlear activity, etc.

In the examples, the neuronal activity, after appropriate filtering, was characterized by the RMS fluctuations of the measured signal, serving as the feedback stimulus. An electric field was subsequently applied in proportion to the RMS. Specifically, the instantaneous RMS activity (e.g., the last 0.25 sec of activity) was low pass filtered with a time constant .tau. to yield A.sub..tau.. This value was compared with a threshold value, as determined by the long time average of the RMS (e.g., the last 30 seconds of activity). The magnitude of the applied field was then derived by calculating the difference between the A.sub..tau. and the threshold multiplied by a gain factor. Any suitable methods and/or algorithm for determining field strength and polarity can be used, e.g., linear and nonlinear proportional feedback, proportional--integral--differential feedback, etc.

The values for instantaneous activity and threshold can be selected empirically, e.g., based on the activity characteristics of the system and the neuronal activity that is to be controlled. The goal is to choose a time scale that distinguishes the activity of interest from the baseline activity of the system. When a timescale for the threshold (e.g., the last 30 seconds of total activity) and instantaneous (e.g., last 0.25 sec of total activity) activity determinations are selected, the difference between such values should permit detection of the onset of the activity of interest.

A gain factor can be chosen such that the output of the applied field is adequate to modulate the neuronal activity that is being monitored. It can be empirically derived, based on previous performance of the neural system and various considerations, including, e.g., magnitude of the onset of the event which is being assessed, magnitude of the applied field necessary to modulate the neural system, characteristics of the field electrodes, characteristics of the neural system environment, etc. In the experiments described herein, a gain was chosen such that a typical difference between A.sub..tau. and the threshold yielded a field in the range of order of 10 mV/mm. Successful control was achieved for the same experiment with gains differing by an order of magnitude indicating that the choice of gain was not critical.

The applied field can utilize the full feedback signal ("full-wave control"), or, it can be half-wave rectified. When half-wave rectification is used, a field is applied only when the instantaneous activity (or the calculated A.sub..tau.) is above (or below) the threshold value. In the examples described below, a field was applied only when there was a positive difference between the instantaneous activity and the threshold. Thus, half-wave rectification indicates that the field is applied in only one direction. For full-wave control, a field is applied continuously when there is any difference between the instantaneous activity (or calculated A.sub..tau.) and the threshold value. The outcome of half-wave rectification is the application of a field in only one direction, while full-wave control results in both negative and positive applied fields, depending upon the sign of the difference between instantaneous activity and threshold. As a result, full-wave control can involve the administration of both excitatory and suppressive signals, while half-wave rectification involves only one kind of signal, either excitatory or suppressive, depending upon the direction of the applied field. The experiments described below show that full-wave control was generally superior to half-wave rectification for seizure suppression, for reducing withdrawal seizures, and for obtaining a more regular baseline of neuronal activity.

Full-wave control may also be desirable to avoid substantial electrode and tissue polarization which occurs when half-wave rectification is used. In the latter case, the electrodes may need to repolarized between field applications, e.g., by applying bias currents to the electrodes.

In general, the duration and intensity of the applied field can be determined by the measured activity. If the purpose is to eliminate neuronal activity, then preferably a field potential, or current, is applied until the activity level is reduced below a threshold level. At this point, the field can be discontinued until activity is observed again. The applied field is preferably not a stationary field, such as the fields described in Gluckman et al., J. Neurophys., 76:4202-4205, 1996; U.S. Pat. No. 5,800,459. See, also, U.S. Pat. Nos. 5,797,965 and 5,522,863.

Activity can also be augmented, induced, or initiated. In the examples, reversing the field potential converted sporadic bursts into a full-blown seizure. In this case, the feedback stimulus is positive feedback, where the applied field is used to enhance activity, e.g., by producing depolarization toward threshold and/or recruiting more neurons into the activity. Here the sign of the gain factor is switched so that a negative field is applied when the RMS activity goes above threshold, forcing the network to become more excitable. The ability to create activity in vitro and in vivo is useful in variety of ways. It can be used to create animal models for epilepsy or electroconvulsive therapy (ECT) and for testing agents which modulate these brain behaviors for therapeutic, prophylactic, and research purposes. It can also be used to induce ECT in humans for therapeutic purposes.

In some instances, a neural system will exhibit ongoing neuronal activity, such as spike activity varying in amplitude and frequency. This information can be processed in any suitable way to serve as a threshold stimulus for the applied field. For instance, the activity in a certain frequency band can be of particular interest because it indicates that certain state of the neural system has been reached, such as epilepsy. It therefore may be desired to apply the electric field only when the system becomes epileptic. This can be accomplished by processing the measured neuronal activity, and applying the field when a predetermined threshold of activity is reached. For example, the long-term average of spontaneous or non-epileptic activity can be determined and used as the stimulus threshold, where no field is applied unless the long-term average, or a function of the average, is exceeded. A particular characteristic of neural activity can also be compared to a matched filter using a temporal, spectral, or wavelet filter, or a nonlinear filter, and its output compared with a threshold.

The methods and devices of the present invention are useful in any endeavor in which it is desired to modify the behavior of a neural system. In general, an applied field in accordance with the present invention can be utilized to modulate any neural activity, including, e.g., synchronized firing, oscillatory firing, pulsating activity, and any in-phase activity of a neural system. Because of such ability to augment or reduce neuronal activity of a neural system, the invention is useful for modulating many kinds of output which arise from neural systems, including motor, sensory, emotional, behavioral, etc.

For example, the methods and devices of the present invention are useful for treating brain diseases characterized by aberrant neuronal activity. Epilepsy, for instance, is a brain disorder characterized by recurrent seizures, affecting 1-2% of the population. In this disease, the pattern of neuronal discharge becomes transiently abnormal. In the examples, an in vitro slice preparation is utilized to illustrate how epilepsy can be treated in accordance with the present invention. When perfused in a high potassium concentration, these networks show a broad range of interictal-like and epileptiform activity, from network wide synchronous events to local and propagating events. Application of the adaptive electric field can be used to suppress the epileptiform activity, effectively treating and controlling the brain disorder.

A modulatory effect can be achieved analogously in situ. For instance, to treat a patient having epilepsy, a device can be utilized which simulates the pair of field electrodes used in the in vitro method. The field electrodes can be positioned in any arrangement which is effective to produce a modulatory field. They can be in contact with brain tissue or associated meninges, e.g., by inserting, through an occipital entrance hole, one, or more, long flat electrode strips that contacts the long axis of the hippocampus surface in the temporal horn of the lateral ventricle. A round electrode (e.g., a single depth electrode with one or more suitable high current contacts) can also be utilized, e.g., by placing it within the long axis of the hippocampus in order to produce a radial electric field. Electrodes can also be external to the brain, e.g., on the scalp. The electrode strip preferably produces an effective electric field. Useful electrode strips include non-polarizing biocompatible electrodes embedded in silastic sheets with sealed electrode-lead connections, similar to those used for cochlear implants, e.g., a Clarion Cochlear Implant, comprising iridium oxide electrodes sealed within a curved silastic silicone elastomer sheath. In another embodiment, a sheet comprising multiple electrodes can be placed over the neocortex in the subdural, subarachnoid, or epidural spaces, or within the sulci of the brain. Thin electrodes can also be inserted into brain tissue. In general, any types or combinations of electrodes, such as those mentioned above, can be used.

In addition to epilepsy, any brain disorder that displays abnormal activity, such as oscillatory or pulsating activity, can be treated analogously. Such diseases, include, schizophrenia, depression (unipolar or bipolar), Parkinson's disease, anxiety, obsessive-compulsive disorder (OCD), etc., where the electric field is applied to the particular brain region exhibiting the abnormal activity, e.g., cortex, hippocampus, thalamus, etc. Parkinson's disease is characterized by decreased activity in cells that produce dopamine. Patients with the disease experience tremors, rigidity, and difficulty in movement. Patients with Parkinson's disease can be treated by applying an electric field in an amount effective to ameliorate one or more symptoms of the disease. Preferably, the applied field is sub-threshold. The field electrodes can be placed in any suitable region of the brain, such as the thalamus or basal ganglia. The electrodes can be of the same in situ type described above for treating epilepsy. The amount of applied field can be changed in response to an electrical activity in the brain, or in response to a manifestation of such electrical activity. For instance, the field can be applied until one or more symptoms are eliminated, such as tremors or difficulty in initiating movement. In such case, the field can be operated manually by the patient, or the behavior can be monitored automatically by feedback sensors either within the brain or placed strategically along the body to sense the behavioral output.

A method of the present invention also relates to restoring or repairing a brain function. These functions include, e.g., sensory functions, such as vision, hearing, smell, touch, and taste, motor activity and function, somatic activity and function, etc. For instance, the method can be useful to treat a condition where an animal (e.g., a human) has lost its vision due to a peripheral defect, such as the loss of an eye, but the visual cortex is largely intact. The present invention can be used to restore vision by creating patterned activity in the brain using an applied field. For example, devices can be used to capture images (e.g., light intensity, wavelength, etc.), process the information, and use the information as a feedback stimulus to the visual cortex, or a subservient pathway, modulating the on-going cortical activity analogously to how epileptic activity was induced from non-epileptic activity as described above and below. Similar strategies can be applied to restoring other lost functions, e.g., hearing or touch to the auditory or somatosensory cortex, respectively.

The present invention also relates to a field-producing device for modifying the neuronal activity of a neural system comprising neurons. Such device is not a voltage-clamp device, or a patch-clamp, as used to clamp the activity of single neurons, or parts thereof. A field-producing device can comprise one or more of the following components: (a) field electrode means for applying an external electric field to a neural system; (b) field application electronic means for generating an external field to a neural system, which is operably connected to (a) field electrode means; (c) measuring means for monitoring the neural activity of the neural system; (d) measurement electronics means for recording neural activity, which is operably connected to (d) measuring electronic means; (e) feedback controller means for determining the amount of external field to apply to the neural system, which is operably connected to (b) field application means and (c) measuring means; (f) sensing means for detecting the external field produced by the field electrode means; (g) sensing electronic means for recording the field produced by the field electrode means, which is operably connected to (f) sensing electrode means and (b) field application means. The device can be used for in vitro applications, or as as in vivo prosthetic devices for treating brain disorders, such as epilepsy and Parkinson's disease, and restoring brain function. In the latter case, the (f) sensing electrodes and (g) electronics are optional.

FIG. 1 illustrates an in vitro field-producing device. In this example, the (b) field application electronic means and (g) sensing electronic means are bundled together, along with an isolation stage. The (d) measuring electronic means is an amplifier of the type typically used to record extracellular and intracellular neuronal activity. The (e) feedback controller means in the example is a computer loaded with the appropriate software for taking data in from the recording electronics and outputting a signal, derived from feedback algorithm, to the field electronics. FIG. 1 also contains a computer ("user interface 7) for recording and displaying information from the various components of the device

The device preferably is for applying a sub-threshold field. It can further comprise a power source for generating the applied field (e.g., a direct or inductive source); external feedback sensors for detecting behavioral output, etc.

For in vivo applications, various methods can be used to place the electrodes the in target tissue, including, visually, stereotactically, endoscopically, ultrasonically, x-rays (such as CT scan), nuclear magnetic resonance, electrical activity, etc.

In addition to identifying characteristics to be used in calculating a feedback stimulus, an additional parameter that can be varied is the choice of the activity that is being measured. Thus, for instance, the feedback stimulus activity can be measured intracellularly from one or more neurons, or extracellularly, capturing field potential from single neurons or a neuronal population. Additionally, the feedback stimulus can be remote or external to the neural system. Thus, the feedback stimulus can be recorded at the site of field application (e.g., using measuring electrodes placed in the tissue), at site remote from the field application, or using a behavioral feedback stimulus, such as movement of a limb when motor activity is modulated, or the ability to experience a sensation when sensory activity is modulated.

The present invention also relates to methods of identifying pharmacological agents which modulate the neuronal activity of a neural system comprising neurons, comprising one or more of the following steps in any order, e.g., measuring the neuronal activity of a neural system; applying an oriented electric field to said neural system effective to modify the neuronal activity of the neural system, wherein the magnitude and polarity of said applied electric field is changed in response to the measured neuronal activity; and administering an agent which modulates the neuronal activity of the neural system. Such a method is especially useful for identifying agents that can be used therapeutically and/or prophylactically in brain disease. Any agent can be administered to the neural system, including, e.g., neurotransmitter agonists and antagonists (such as, serotonin, dopamine, GABA, glutamate), sympathomimetics, cholinergics, adrenergics, muscarinics, antispasmodics, hormones, peptides, genes (sense and antisense, including genetic therapy), metabolites, cells (e.g., where neural grafting is being used as a modulatory therapy), sedatives, hypnotics, anti-epileptics (e.g., acetazolamide, amphetamine, carbamazepine, chloropromazine, clorazepate, dextroamphetamine, dimenhydrinate, ephedrine, divalproex, ethosuximide, magnesium sulfate, mephenytoin, metharbital, methsuximide, oxazepam, paraldehyde, pamethadione, phenacemide, phenobarbital, phensuximide, phenytoin, primidone, trimethadione, valproate, etc.), hormones, peptides, etc.

In an in vitro method and device of the present invention, a slice of rat brain tissue obtained from the hippocampus of the temporal lobe is perfused with an oxygenated physiological perfusate fluid ("ACSF" or artificial cerebrospinal fluid) in an interface-type perfusion chamber (e.g., Hass-style) comprising an inlet 9 and outlet 10 for continuously replacing the perfusate. A heated oxygen/carbon dioxide gas (95% oxygen, 5% carbon dioxide at 35.degree. C.) is provided through inlet 11. The top of the chamber can be open, or covered.

The anatomy of the brain tissue includes layers of pyramidal neurons of the Cornu Ammonis (CA) regions. In order to induce seizures, the ACSF perfusate is replaced through the inlet 9 with a high potassium solution, comprising 8.5 mM potassium and 141 mM chloride. The elevated potassium produces epileptic activity characterized by events in the form of spontaneous burst firings and seizure-like events within the two regions (CA3 and CA1 respectively) at opposite ends of the Cornu Ammonis. Seizure-like activity can also be produced by other treatments, including, penicillin, low magnesium, kainic acid lesions, or any one of the epileptogenic compounds. Additionally, naturally-occurring and induced mutants which result in aberrant brain activity, including mutants produced by genetic-engineering, e.g., in channel genes and receptor genes, can be used as a source of brain tissue.

The brain tissue slice labeled by reference numeral 1 in FIG. 1 is supported on a nylon mesh 2 submerged in artificial cerebrospinal fluid the perfusate within a chamber formed by an annular wall 3. A pair of parallel spaced Ag--AgCl field electrode plates 4 (F1, F2) are placed on the floor of the chamber, positioned in such a manner to produce an electric field parallel to the soma-dendritic axis. The field electrodes 4 are spaced apart from each other, for example by 1.8 cm. An electric field is established between the electrodes 4 in the perfusion chamber within which the tissue slice 1 is submerged in the perfusate fluid. A pair of ground electrodes 10 (G) are positioned on the floor of the chamber. A pair of Ag--AgCl sensing electrodes 5 (S1, S2), placed 12 mm apart, are shown in FIG. 1 for sensing the field produced by the electrodes 4 and to feedback control the field in the chamber. Micropipette measuring electrodes 12 (above the chamber) are used to measure neuronal activity extracellularly. The electronics are set up so that the potential between S1 and S2 is equal to a gain (of 1 or 0.1) times the program potential (from the computer or a waveform generator).

The measuring electrodes 12 are adjacent to the pyramidal cell layer of the brain tissue slice 1 at a position along a field isopotential to minimize recording artifact by means of differential amplification. Such positional arrangement of the electrodes 12 allows for continuous recording of neuronal activity in the brain tissue slice 1 despite relatively substantial changes in the electric field established between the electrodes 4.

The potential measured through the measuring electrodes 12 are filtered through the recording amplifier 6 and directed to the user interface for monitoring and parameter control 7 and the feedback controller 8. The monitoring and parameter control 7 can accept input from the recording electrode 6 and the feedback controller 8, and display and record such input. Based on the measured activity from the recording electrodes 12, an electric field is externally imposed on the brain tissue slice 1 by applying a potential difference to the electrodes 4 through the field application electronics 9. The amount of generated field is determined by the feedback controller 8 which accepts information from the recording (measuring) electrode electronics 6 about the activity of the neural system, and using a selected algorithm (either as software, hardware, or a combination), generates a signal to the field electronics 9. This signal to the field electronics 9 results in the application of a field by the field electrode means 4. The field application electronics 9 comprises an amplifier circuit through a 4-probe feedback technique which applies a potential (or current) between the field electrodes 4 in order to set the field between the sensing electrodes 5 equal to the amplifier's program voltage times a gain (gain=1 or 0.1). Built into this circuit is a layer of ground isolation stage that allow its potentials to float from those of the recording system.

The electronics used to control the field can comprise an input stage A, a standard summing amplifier with a switchable gain of either 1.0 or 0.1 and a low pass frequency of 10 kHz. The output of A is sent both to a monitoring stage B, and to an isolated output stage C. The monitoring stage B can be composed of a unity gain non-inverting amplifier which acts as a buffer to a monitoring channel for recording the summed input. The output stage C can be a circuit utilizing the Analog Devices AMP01 instrumentation amplifier and a OP37 op-amp which provides the feedback stabilized field via the Ag--AgCl electrode plates in a chamber D. This stage can be separately powered by rechargeable batteries in order to isolate this circuit from measurement ground. Unity gain buffers (e.g., from an AD712 op-amp) used to minimize the current through sensing plates S1 and S2.