Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2014 Mar 1.
Published in final edited form as: Neurosurg Focus. 2009 Jul;27(1):E8. doi: 10.3171/2009.4.FOCUS0983

In vivo performance of a microelectrode neural probe with integrated drug delivery

Pratik Rohatgi 1,2, Nicholas B Langhals 2, Daryl R Kipke 2, Parag G Patil 1,2,3
PMCID: PMC3938951  NIHMSID: NIHMS233678  PMID: 19569896

Abstract

Object

The availability of sophisticated neural probes is a key prerequisite in the development of future brain machine interfaces (BMI). In this study, we developed and validated a neural probe design capable of simultaneous drug delivery and electrophysiology recordings in vivo. Focal drug delivery has promise to dramatically extend the recording lives of neural probes, a limiting factor to clinical adoption of BMI technology.

Methods

To form the multifunctional neural probe, we affixed a 16-channel microfabricated silicon electrode array to a fused silica catheter. Three experiments were conducted to characterize the performance of the device. Experiment 1 examines cellular damage from probe insertion and the drug distribution in tissue. Experiment 2 measures the effects of saline infusions delivered through the probe on concurrent electrophysiology. Experiment 3 demonstrates that a physiologically relevant amount of drug can be delivered in a controlled fashion. For these experiments, Hoechst and propidium iodide were used to assess insertion trauma and the tissue distribution of the infusate. Artificial cerebral spinal fluid and tetrodotoxin were injected to determine the efficacy of drug delivery.

Results

The newly developed multifunctional neural probes were successfully inserted into rat cortex and were able to deliver fluids and drugs that resulted in the expected electrophysiological and histological responses. The damage from insertion of the device into brain tissue was substantially less than the volume of drug dispersion in tissue. Electrophysiological activity, including both individual spikes as well as local field potentials, was successfully recorded with this device during real-time drug delivery. No significant changes were seen in response to delivery of artificial cerebral spinal fluid as a control experiment, whereas delivery of tetrodotoxin produced the expected result of suppressing all spiking activity in the vicinity of the catheter outlet.

Conclusions

Multifunctional neural probes such as the ones developed and validated within this study have great potential to help further understand the design space and criteria for the next generation of neural probe technology. By incorporating integrated drug delivery functionality into the probes, new treatment options for neurological disorders and regenerative neural interfaces utilizing localized and feedback controlled delivery of drugs can be realized in the near future.

Keywords: Drug delivery, neural engineering, microelectrode array

Brain machine interfaces (BMIs) depend on technology that detects neural activity in the central nervous system. Neural firing patterns are often recorded from microelectrodes, which in turn are used to measure an output signal that can be manipulated by a patient’s brain activity 4, 10, 11, 41. A new avenue to interface with the central nervous system can be achieved by providing drug delivery capabilities through existing neural probe technology.

A large barrier to the clinical adoption of BMI technologies is electrode failure. Device failure is hypothesized to result from a foreign-body response that gradually encapsulates the BMI implant with scar tissue 17, 30, 33. This process increases the distance between electrode sites and neuronal sources, reducing the ability to record brain activity. Systemic corticosteroid delivery has been investigated to mitigate this inflammatory process; however, medication side effects limit the duration and strength of the therapy 38. Systemic drug delivery requires large administered doses in order to penetrate the blood brain barrier at a sufficient concentration to have a measurable effect in the brain. Consequently, there is a great interest in controlled and targeted drug delivery to the brain so that small localized doses can be used instead of large systemic drug doses. Because the reactive tissue responses against implanted neural prostheses are focal by nature, it may be fruitful to use focal interventional strategies to mitigate this process.

As nervous system processes are driven by both chemical and electrical signals, multimodal therapies may have applications in the future. In fact, systemic drug delivery is still the principal treatment for many neurological disorders. For example, Parkinson’s disease treatment might be advanced by adding a drug delivery capability to deep brain stimulating electrodes, potentially reducing the need for systemic medications and the associated side effects. Epilepsy treatment devices may be envisioned to provide localized drug delivery therapy when triggered by electrophysiological conditions suggestive of an oncoming seizure. Robust investigational multifunctional neural probes are needed for the research required to realize these types of therapeutic devices.

One method to provide localized drug delivery to the central nervous system is through convection enhanced-delivery (CED) 2, 20. CED is the process of controlled infusion directly into brain parenchyma driven by a pressure source or pump. The volume of distribution into the tissue depends only upon the injection volume. Furthermore, the concentration profile of the infusate is homogeneous within the volume of distribution and independent of infusion rate, duration, and volume 13. Over time the concentration gradient between the volume of distribution and the remaining brain parenchyma further disperses the infusate by diffusion. Yet, the high degree of control afforded by CED methods is typically preferable to primarily diffusion driven drug delivery, to match the high spatial resolution of microelectrode technology to the fluidic performance.

Several attempts have been made to integrate microfluidic drug delivery channels into microelectrode arrays 5, 19, 30, 32, 35, 36. These devices have the same microelectrode features used in BMI research with the added ability to provide CED-driven drug delivery. However, much of this work has focused on the technical aspects of the fabrication process rather than the physiological performance of the device, due to the extremely high technical requirements needed to develop these technologies. Consequently, a simplified neural drug delivery probe ready for in vivo experimentation is needed for the data-driven development of more advanced microfabricated devices.

The objective of this work is to develop and to validate a neural probe design capable of simultaneous drug delivery and electrophysiological recordings in vivo. The device was designed to simplify convection-enhanced drug delivery for acute microelectrode array experiments. Proof-of-concept animal experiments were performed to assess drug dispersion and tissue damage, to determine the volume-effects of infusion on neural activity, and to modulate neural activity through drug infusion. The results of these experiments allow us to determine if measured changes in neural activity are due to the direct actions of the infused drug or by non-specific secondary effects. These proof-of-concept experiments also demonstrate the increased functionality of concurrent electrophysiological and focal drug delivery, which not only monitor neural activity but also interface with the chemical environment, leading to greater understanding how BMI devices function and expand the potential BMI design space for clinical applications.

Methods and Materials

Multifunctional neural probe description

Multifunctional neural probes were created by incorporating a drug delivery catheter to a chronic-style Michigan microelectrode array on an acute experimental package (Fig. 1). Microelectrode arrays were fabricated using standard silicon processing techniques common for Michigan neural probes 18. The number of microelectrode sites and their spacings were selected to uniformly sample across the depth of rat cortex, and these dimensions are widely used in Michigan neural probe technology. This microelectrode design incorporated a flexible cable upstream to the implanted portion of the microelectrode array, which allowed the array to bend and mate to the drug-delivery catheter without breaking 16. The drug-delivery catheter and pump-connection adapters were bonded to the back of the device. This design was manufactured by NeuroNexus Technologies, Ann Arbor MI.

Figure 1.

Figure 1

Open in a new tab

Top, the full neural probe and attached fluidic line. Bottom, a magnified view of the electrode sites on the array. Each of the 16 iridium microelectrode sites have a surface area of 703 μm2 and are spaced 100 μm apart. The catheter is made of fused silica covered in a polyimide sheath with an outer diameter of 165 μm and an inner diameter of 127 μm.

Supporting equipment

Fluids were delivered through the catheter using a 10 μl syringe (1700 Series, Hamilton Company) driven by a microsyringe pump (UltraMicroPump 4, World Precision Instruments). For a 1 μl injection, the percent error in delivery volume was 0.133% (0.999 μl ± 21 nl, n=10). A programmable linear actuator was used (M-230.25, Physik Instrumente) to insert the probe into the brain. Automated implantation was preferred over manual insertion in order to minimize tissue damage 1.

Animal preparation and set-up

A total of twenty Sprague-Dawley rats (300-600 g) were used for these experiments following an approved protocol (#08227) in accordance with the University Committee on Use and Care of Animals guidelines at the University of Michigan. The animals were housed and fed by university husbandry staff. All animal experiments involved non-survival surgery. Urethane (1.25 g/kg body weight) was used to provide prolonged anesthesia to the animal for the duration of the experiment. Urethane is commonly used because it minimally affects neurotransmitter levels involved in network oscillations in the brain 14, 24, 25. After the animal was secured in the stereotactic frame, craniotomies were opened bilaterally over primary visual cortex, located 6.5 mm posterior to bregma and 3 mm lateral to midline. The dura was resected to prevent the probe from buckling or breaking during insertion. After the surgery, the animal was placed in a faraday cage to minimize electrical noise that often obscures spike activity. The linear actuator, microsyringe pump, and electrophysiological recording amplifiers were then added to the set-up (Fig. 2). Supporting instrumentation was powered down to minimize noise during each recording session. A flashing LED array with 1 second on and 5 seconds off cycle was placed in front of the animal in order to increase the probability of measuring driven action potentials in the visual cortex.

Figure 2.

Figure 2

Open in a new tab

The animal experimental setup with the neural probe inserted into rat cortex. The animal is secured in a stereotactic frame. The faraday cage and electronic amplifiers required for electrophysiological recordings are not shown.

Experimental procedure

The device was prefilled with the infusate before insertion so air would not be injected into the brain. The solution was not warmed to body temperature because it would return to room temperature prior to infusion due to the small volume used (< 100 μl) and required setup time (20-30 min). The probe was positioned over the craniotomy touching the surface of the brain using the stereotactic manipulator and inserted with the linear actuator to an initial depth of 2 mm at a rate of 1.2 mm/sec. If no action potentials were identified at that depth, the probe was either advanced or withdrawn until spiking activity was measured. The catheter outlet placement was limited to a range 1.5-5 mm below the surface of the cortex. Only a single 1 μl infusion at a rate of 100 nl/min was delivered into each craniotomy for any given trial. These parameters were selected based on prior benchtop work briefly summarized in the results. Electrophysiology was typically recorded before, during, and after drug infusion.

Electrophysiological recordings

For all animals in this study, electrophysiological data was acquired using a TDT Pentusa Recording System (Tucker-Davis Technologies). These neuronal signals were acquired through the 16-channel electrodes with a head-stage buffer amplifier to avoid signal loss in data transmission. Signals were sequentially filtered by an anti-aliasing filter in the preamplifier, digitized at ~25 kHz sampling rate, and digitally band-pass filtered from 2-5000 Hz. Wideband signals were acquired to capture both spiking and local field potential activity. Signals were continuously recorded in segments ranging from 30 seconds to over 10 minutes in duration. To further reduce noise in the recordings, common average referencing (CAR) was used in real-time 22. This technique calculates a reference signal by averaging data on every channel and subtracts this value from each electrode recording to eliminate noise that is correlated across the array. CAR is necessary for real-time drug delivery recordings because the active syringe pump is a large source of electrical noise. Neuronal activity was better assessed during experimentation by performing this technique in real-time and helped guide probe insertion depth.

In vivo experiments

Three different animal experiments were conducted to evaluate device function in vivo. The purpose of the first experiment was to assess drug dispersion and tissue damage by imaging infused stains. The infusate consisted of a solution containing Hoechst 33342 (Invitrogen) and propidium iodide (PI) (Invitrogen) each at 1 mg/ml in artificial cerebral spinal fluid (aCSF) (Harvard Apparatus). The purpose of the second experiment was to determine if the volume-effects of the infusate would alter measured neural activity. For this experiment, 1 μl aCSF was delivered through the device into brain tissue. The purpose of the third experiment was to observe changes in neuronal activity due to a delivered drug. For this experiment, tetrodotoxin (Tocris Bioscience) at 10 ng/μl was added to the solution used in the first experiment and delivered through the device into brain tissue.

Histology

After each experiment that required tissue assessment, the animal was perfused with 4% paraformaldehyde for histology. After perfusion, the brain was excised from the skull and refrigerated at 4° C in 4% paraformaldehyde for at least 24 hrs. A vibratome (myNeuroLab) was used to make 100 μm thick transverse brain slices that were mounted to slides using Prolong Gold Antifade Reagent (Invitrogen). After allowing the slides to fix for at least 24 hours, the tissue slices were imaged using an epifluorescence wide-field microscope (MZFLIII, Leica). A 12-bit color CCD (CoolSNAP PRO Color, MediaCybernetics) mounted to the microscope was used to digitally store each image. Only tissue slices with dye visible by eye through the microscope were imaged. The exposure time for each image was adjusted such that the maximum signal intensity matched the saturation point of the CCD. The image background was normalized in post-processing software (Photoshop, Adobe Systems).

Electrophysiological analysis

Neural recording data was analyzed after acquisition using custom automated MATLAB (Mathworks Inc., MA) software, as described in detail elsewhere22, 23. As an overview, the wide-band recordings were filtered in software to isolate the spike data (300-5000 Hz) from the local field potential data (1-100 Hz). To identify individual units, the high frequency data was thesholded using a window set at 3.5 standard deviations below the mean of the data. A 2.4 ms waveform was extracted from the data stream at each threshold crossing. To group isolated waveforms to a single neuronal unit, principal component analysis was then completed and the resultant components were separated into individual clusters using Fuzzy C-means clustering. The results of this spike sorting process from a data set within this study are shown in Figure 3.

Figure 3.

Figure 3

Open in a new tab

An example of neural spikes recorded simultaneously on 4 of the 16 microelectrode sites of the device. Depth from the surface of the cortex is indicated for each channel. Two channels show two distinct patterns of unit activity as determined by the spike sorting algorithm described in the methods.

Local field potentials were primarily analyzed in the frequency domain. Power spectrums and spectrograms were calculated using the signal processing toolbox within MATLAB. For a comparative analysis, data for all channels of a given trial were combined and the LFP power was averaged to create delta (0-3 Hz), theta (3-7 Hz), alpha (8-12 Hz), beta (12-30 Hz), and gamma bands (30-100 Hz). The mean value for each band before and approximately 1 hour after infusion was compared using a paired t-test.

Results

This study was performed using the multimodal neural probe illustrated in Figure 1 and setup shown in Figure 2 as described in the methods. Figure 3 demonstrates the quality of units recorded with the device and the effectiveness of the spike sorting algorithms.

Validation and optimization of drug delivery parameters

Prior work was performed to validate the probe design and performance in 0.6% agarose gel in order to determine appropriate probe insertion and drug infusion rates. This concentration of agarose has similar poroelastic properties to brain parenchyma with respect to volume of delivery, infusion pressure, and insertion force 8. Results (not shown) from the agarose tests showed that slower infusion rates decreased backflow, permitting the catheter outlet to act more like a point-source. Additionally, faster probe insertion speeds minimized agarose damage by cutting sharply into the gel rather than creating large deformations and tears. Damage disrupts the tissue-device interface, creating a low-resistance fluid shunt along the insertion path. This allows the infusate to reflux outside of the cortex instead of penetrating into the brain parenchyma. These results are consistent with findings from other studies on the optimization of convection-enhanced delivery parameters in cortex 6, 27. These studies demonstrate that backflow is significantly increased for infusion rates above 1.0 μl/min versus infusion rates below 0.5 μl/min. They also show that larger cannula sizes increase backflow, but tissue relaxation time after device insertion did not affect backflow. Consequently, all animal experiments for this study were performed using an infusion rate of 100 nl/min and insertion speed of 1.2 mm/sec, which was the fastest speed of the linear actuator.

Assessment of drug dispersion and tissue damage

Hoechst and propidium iodide (PI) stains were successfully visualized in brain tissue after histology (Fig. 4). Hoechst and PI both intercalate with DNA, but PI is membrane impermeable and thus only stains cells with damaged cellular membranes 36. Consequently, tissue stained with Hoechst reflects drug dispersion within the brain whereas PI only highlights areas of tissue damage. Hoechst was used as a visible proxy for TTX distribution in tissue because both are hydrophilic and have molecular weights on the same order of magnitude. Hoechst dispersed in a roughly spherical distribution through the tissue elongated along the axis of the probe, indicating that the infusate penetrated into brain parenchyma rather than primarily shunting outside the brain along the insertion tract. Further, the dye did not stain slices extending to the surface of the brain. The area of tissue stained with PI was tightly focused around the probe insertion tract and markedly reduced from the area stained with Hoechst. Therefore, the volume of tissue damaged by the device was less than the volume of drug distribution.

Figure 4.

Figure 4

Open in a new tab

Histology of a single injection site of Hoechst (upper panel) and PI (lower panel) imaged using an epifluorescence wide-field microscope at 50x. The brain tissue was sectioned in transverse slices relative to the long-axis of the neural device. Depth is labeled with respect to the first tissue slice with visible dye. The same tissue slice is shown in A and B for a given depth. The device outlet location cannot be localized to a specific tissue slice. Some images show artifact due to the slide fixation process.

Effects of artificial cerebral spinal fluid and tetrodotoxin during infusion

Real-time electrophysiological data during artificial cerebral spinal fluid (aCSF) and tetrodotoxin (TTX) delivery was successfully recorded, demonstrating the ability of the device to simultaneously record electrophysiology and deliver drugs. Infusion started approximately 30 seconds into the recording window and continued for 600 seconds. Representative data from single aCSF and TTX infusions are shown in Figure 5 for electrode sites with identified units. Only a single unit is illustrated for channels with multiple units. Each electrode site is represented by a spectrogram and raster plot. The spectrograms show the local field potential (LFP) power as a function of time and frequency. A raster plot is placed above each spectrogram and displays a single unit firing event as a vertical line. The line density reflects the firing rate of each neuron.

Figure 5.

Figure 5

Open in a new tab

LFP spectrogram and raster plots during the delivery of aCSF (A) and TTX (B) for sites with identified units. The start of drug delivery is marked by a triangle on the x-axis, which begins roughly 30 sec into the recording block and continues for 600 sec. Each site is labeled in relation to its distance from the catheter outlet. Each raster plot shows every 10th identified spike. Power is presented on the same scale for all sites of a given trial and hot (red) colors indicate higher power and cold (blue) colors indicate lower power. The spectrograms for microelectrode sites without identified units were similar to channels with identified units for both aCSF and TTX delivery (not shown). Note the deactivation of neural activity during TTX delivery at approximately 400 sec.

TTX strongly inhibits voltage-gated sodium channels, thus stopping action potential propagation along all neuronal axons 28, 29. Neuroscientists use TTX to create reversible lesions in order to better study nervous system function 3, 12, 26, 34, 40, 42. TTX was selected as the delivery drug because it changes neuronal activity in a manner that is straightforward to interpret. Neither the LFP power nor the spiking activity changed during the aCSF infusion, but both measures dramatically changed during TTX delivery. All spiking activity stopped approximately 400 sec into the TTX infusion, which corresponds to a delivered volume near 0.66 μl. LFP power dropped at this time point and continued to do so throughout the remainder of the recording window. Since aCSF did not alter unit activity, changes in activity can be attributed to drug delivery.

Spike data before and after infusion

Spike data was compared before aCSF and TTX delivery and 1 hour after delivery. Spikes were identified before and after aCSF delivery. Units showed similar waveform shape at a given microelectrode site, suggesting that the same neuron was active during both recordings segments (Fig. 6). No spikes were identified after TTX delivery. This is consistent with results showing that spiking activity stopped during TTX infusion. TTX reversibly inhibits neuronal activity, however the recovery time course for a 1 μl injection at 10 ng/μl can be as long as 20-24 hours 42. Consequently, the return of neuronal activity was neither observed nor expected one hour post TTX delivery.

Figure 6.

Figure 6

Open in a new tab

The morphology of a unit before and after aCSF delivery. In this example, a second unit was identified after aCSF delivery.

Local field potential data before and after infusion

LFP data were also compared before and 1 hour after aCSF and TTX delivery. Data for two different trials of aCSF and TTX infusions were aggregated to statistically compare a change in LFP power before and after infusion using a paired t-test (Table 1). The aCSF infusion did not create a statistically significant difference in power for any frequency band. Conversely, TTX infusion created a statistically significant (p < 0.05) change in power in every frequency band. Changes in LFP were not obviously different at microelectrode sites closer to the catheter opening (Fig. 7).

Table 1.

Comparison of LFP power before and approximately 1 hour after aCSF and TTX delivery.

Frequency Band aCSF P-Value TTX P-Value
Delta (0-3 Hz) 0.419 < 0.05
Theta (3-7 Hz) 0.763 < 0.05
Alpha (8-12Hz) 0.580 < 0.05
Beta (12-30 Hz) 0.448 < 0.05
Gamma (30-100 Hz) 0.298 < 0.05
Open in a new tab

P-values were calculated using a paired t-test. Data was pooled from two different injections to provide a sample size of 32 for each statistical comparison.

Figure 7.

Figure 7

Open in a new tab

Changes in LFP power before (blue) and approximately 1 hour after (yellow) aCSF (A) and TTX (B) infusion. Each site is labeled in relation to its distance from the catheter outlet. 60 Hz electrical noise is evident on several electrode sites. Note that TTX delivery caused an appreciable decrease in power at every site, whereas aCSF delivery caused no notable change in LFP power.

Discussion

The multifunctional neural probe used in this study successfully modified neuronal activity through drug delivery as measured by electrophysiology. However, the drug delivery resolution for the parameters studied is less than electrophysiological resolution of the device. This is reflected by the uniform change in electrophysiology measured in response to drug delivery on every site of the microelectrode array. Tighter controlled drug delivery is needed to compare electrophysiological recordings from sites measuring neural activity modulated by the drug to other sites on the array measuring unaltered neural activity during the same experiment. Alternatively, a device with a fewer number of or further spaced microelectrode sites can be used in order to avoid taking recordings that do not provide additional information about drug kinetics, delivery, or neuronal modulation.

Assessment of drug dispersion and tissue damage

The method to visualize drug delivery demonstrates that the infusate penetrated into the brain tissue with negligible backflow. However, the elliptical drug distribution pattern suggests that the catheter outlet could not be idealized as a point-source within this study. To better plan and carry out future experiments, a model is needed to correlate the injection volume with the volume of distribution in tissue. Biotransport in the brain is complicated by various processes including diffusion, bulk flow, molecular uptake, metabolism, and the device-tissue interface 31, 39. Several efforts have been made to mathematically describe drug distribution for convection-enhanced delivery and can be applied to drug delivery with this device as well 7, 21, 27.

Tissue damage was evident with propidium iodide. It is not understood if the tissue damage around the probe is cross-sectional area dependant and if larger catheter sizes decrease the ability to measure units. The multifunctional electrode had a crosssectional area of over 20,000 μm2 which is dramatically larger than the typical crosssectional area for a Michigan microelectrode array of about 1500 μm2 15. Although no quantitative statistical analyses were performed for comparison, electrophysiological performance for the multifunctional device was comparable to the Michigan microelectrode array despite the larger size based on experience within our laboratory. A better understanding of tissue damage would help steer device design and help determine if the smaller-sized and technically more involved microfabricated fluidic channels are necessary to improve multifunctional device performance.

Volume-effects of infusion on neural activity

Because neuronal activity was not changed with aCSF delivery, potential tissue damage due to fluid delivery may not be physiologically significant as assessed by electrophysiology for the parameters studied (1 μl at 100 nl/min). In fact, the upper limit of pressure injections into cortex before tissue damage is thought to range from 12-20 μl 9. The trauma of the device insertion did not completely lesion the surrounding neuronal tissue as evident by the ability to record spikes prior to drug delivery. It is not clear whether tissue damage was due solely to probe insertion or partially dependant on fluid delivery. Possible neuronal displacement or changes in intracranial pressure may not be physiologically relevant. Because the dura is open, the infusate can shunt or the brain can herniate outside of the craniotomy to dissipate pressure and allow tissue around the probe to relax. This area requires more investigation because chronic drug delivery devices will require the craniotomy to be sealed from the outside environment.

Drug induced modulation of neural activity

The infusion of TTX successfully suppressed neuronal activity in the vicinity of the neural probe. TTX suppressed all spiking activity during infusion and lowered local field potential power. These changes started roughly two-thirds of the way through the infusion at every electrode depth, indicating that tissue around each electrode site was exposed to TTX in the same timeframe. The tissue at the catheter outlet could be acting like a valve that does not open until sufficient backpressure is created. Backflow may also act to first shunt TTX along the probe tract and suppress neuronal activity immediately next to the probe at a faster time course than drug penetration into brain parenchyma due to a difference in flow resistivity. Additionally, because spikes are typically recorded from a distance of up to 50 μm from the electrode array, this signal would first be suppressed by the backflow of TTX. LFPs are recorded from a radius of several hundred microns away from the electrode array, and thus changes in this signal may be more indicative of infusate penetration into brain tissue and drug kinetics. If the catheter outlet acted as an ideal point source with appropriately selected drug delivery parameters, changes in neuronal activity should first be measured at electrode sites most proximal to the outlet and propagate to distal sites due to the high spatial resolution of the array. The delivery rate for these experiments was selected to minimize backflow of infusate to the surface of the cortex, but an even tighter control of backflow is needed to see a progressive change of neuronal modulation across the electrode sites. Consequently, an even slower infusion rate should be used to further minimize backflow and backpressure effects by allowing the infusate to penetrate into tissue, thus allowing the outlet to behave more closely to an ideal point source.

Measured action potentials were only suppressed during infusion, thus indicating that drug diffusion from the outlet did not substantially change the activity of neurons in the vicinity of the microelectrode array. Yet, diffusion may limit the use of this device in studies that require very tight control of drug delivery. Devices that rely on diffusion-based drug delivery create a volume of distribution extending 400 μm radially, which is larger than the typical microelectrode site spacing 37. Microfabricated devices have been created with valves that minimize the potential effects of diffusion, but their effectiveness has not been evaluated in vivo with respect to electrophysiological measures 32.

Conclusions

This device provides a platform to drive further developments in drug delivery in order to manufacture clinically useful multifunctional neural prostheses. As brain machine interface technology matures, drug delivery can provide an additional means to interface with the central nervous system to either improve system performance or provide therapy. Additionally, the ability of the device to monitor electrophysiological changes in real-time in response to drug delivery will become a powerful investigational tool in neuroscience and pharmacology.

Studies to investigate different device designs including varied catheter size, outlet geometries, and delivery parameters should be used to further drive microfabricated multifunctional probe design. Additionally, polymer-based microelectrode array technology should soon replace the use of silicon substrate electrodes. This change will improve device usability by allowing the catheter to bend without breaking the bonded microelectrode array. Efforts are also now actively underway to modify the device to create a fully implantable chronic drug delivery system.

Acknowledgments

The authors would like to acknowledge the financial support of the Center for Neural Communication Technology (NIBIB, P41 EB002030) and the University of Michigan Medical School Student Biomedical Research Program. The authors would like to thank Dr. Kip Ludwig for project planning and manuscript editing assistance, Denzel Davis, Matthew Gibson, and the other members of the Neural Engineering Laboratory of the University of Michigan for their assistance with this study.

Funding: This research was supported by the Center for Neural Communication Technology, a P41 Resource Center funded by the National Institute of Biomedical Imaging and Bioengineering (NIBIB, P41 EB002030) and supported by the National Institutes of Health (NIH). Pratik Rohatgi received specific support by the Student Biomedical Research Program funded through the University of Michigan Medical School.

Footnotes

Disclosures

Dr. Daryl R. Kipke has a significant financial interest in NeuroNexus Technologies. Nicholas B. Langhals is a part-time consultant at NeuroNexus Technologies. Dr. Parag G. Patil and Pratik Rohatgi have no conflicts to disclose.

Portions of this work were presented in poster form at the 2008 Fall Student Biomedical Research Forum and 2008 Engineering Graduate Symposium, both held in Ann Arbor, MI in November 2008.

References

  • 1.Bjornsson C, Oh S, Al-Kofahi Y, Lim Y, Smith K, Turner J, et al. Effects of insertion conditions on tissue strain and vascular damage during neuroprosthetic device insertion. Journal of Neural Engineering. 2006;3:196–207. doi: 10.1088/1741-2560/3/3/002. [DOI] [PubMed] [Google Scholar]
  • 2.Bobo RH, Laske DW, Akbasak A, Morrison PF, Dedrick RL, Oldfield EH. Convection-enhanced delivery of macromolecules in the brain. Proc Natl Acad Sci U S A. 1994;91:2076–2080. doi: 10.1073/pnas.91.6.2076. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Boehnke SE, Rasmusson DD. Time course and effective spread of lidocaine and tetrodotoxin delivered via microdialysis: An electrophysiological study in cerebral cortex. Journal of Neuroscience Methods. 2001;105:133–141. doi: 10.1016/s0165-0270(00)00348-4. [DOI] [PubMed] [Google Scholar]
  • 4.Chapin JK, Moxon KA, Markowitz RS, Nicolelis MAL. Real-time control of a robot arm using simultaneously recorded neurons in the motor cortex. Nat Neurosci. 1999;2:664–670. doi: 10.1038/10223. [DOI] [PubMed] [Google Scholar]
  • 5.Chen J, Wise KD, Hetke JF, Bledsoe SC., Jr A multichannel neural probe for selective chemical delivery at the cellular level. IEEE Trans Biomed Eng. 1997;44:760–769. doi: 10.1109/10.605435. [DOI] [PubMed] [Google Scholar]
  • 6.Chen MY, Lonser RR, Morrison PF, Governale LS, Oldfield EH. Variables affecting convection-enhanced delivery to the striatum: A systematic examination of rate of infusion, cannula size, infusate concentration, and tissue-cannula sealing time. J Neurosurg. 1999;90:315–320. doi: 10.3171/jns.1999.90.2.0315. [DOI] [PubMed] [Google Scholar]
  • 7.Chen ZJ, Broaddus WC, Viswanathan RR, Raghavan R, Gillies GT. Intraparenchymal drug delivery via positive-pressure infusion: Experimental and modeling studies of poroelasticity in brain phantom gels. IEEE Trans Biomed Eng. 2002;49:85–96. doi: 10.1109/10.979348. [DOI] [PubMed] [Google Scholar]
  • 8.Chen ZJ, Gillies GT, Broaddus WC, Prabhu SS, Fillmore H, Mitchell RM, et al. A realistic brain tissue phantom for intraparenchymal infusion studies. J Neurosurg. 2004;101:314–322. doi: 10.3171/jns.2004.101.2.0314. [DOI] [PubMed] [Google Scholar]
  • 9.Demer J, Robinson D. Effects of reversible lesions and stimulation of olivocerebellar system on vestibuloocular reflex plasticity. J Neurophysiol. 1982;47:1084–1107. doi: 10.1152/jn.1982.47.6.1084. [DOI] [PubMed] [Google Scholar]
  • 10.Donoghue JP. Connecting cortex to machines: Recent advances in brain interfaces. Nat Neurosci. 2002;5(Suppl):1085–1088. doi: 10.1038/nn947. [DOI] [PubMed] [Google Scholar]
  • 11.Fitzsimmons NA, Drake W, Hanson TL, Lebedev MA, Nicolelis MA. Primate reaching cued by multichannel spatiotemporal cortical microstimulation. J Neurosci. 2007;27:5593–5602. doi: 10.1523/JNEUROSCI.5297-06.2007. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Gallo M, Bielavska E, Roldán G, Bures J. Tetrodotoxin inactivation of the gustatory cortex disrupts the effect of the N-methyl--aspartate antagonist ketamine on latent inhibition of conditioned taste aversion in rats. Neuroscience Letters. 1998;240:61–64. doi: 10.1016/s0304-3940(97)00897-5. [DOI] [PubMed] [Google Scholar]
  • 13.Groothuis DR. The blood-brain and blood-tumor barriers: A review of strategies for increasing drug delivery. Neuro-Oncology. 2000;2:45–59. doi: 10.1093/neuonc/2.1.45. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Hara K, Harris RA. The anesthetic mechanism of urethane: The effects on neurotransmitter-gated ion channels. Anesthesia & Analgesia. 2002;94:313–318. doi: 10.1097/00000539-200202000-00015. [DOI] [PubMed] [Google Scholar]
  • 15.Hetke J, Anderson D, Wise K. Design ranges for silicon multichannel neural probes. 1 [Google Scholar]
  • 16.Hetke J, Lund J, Najafi K, Wise K, Anderson D. Silicon ribbon cables for chronically implantable microelectrode arrays. IEEE Transactions on Biomedical Engineering. 1994;41:314–321. doi: 10.1109/10.284959. [DOI] [PubMed] [Google Scholar]
  • 17.Kipke DR, Shain W, Buzsaki G, Fetz E, Henderson JM, Hetke JF, et al. Advanced neurotechnologies for chronic neural interfaces: New horizons and clinical opportunities. Journal of Neuroscience. 2008;28:11830. doi: 10.1523/JNEUROSCI.3879-08.2008. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Kipke DR, Vetter RJ, Williams JC, Hetke JF. Silicon-substrate intracortical microelectrode arrays for long-term recording of neuronal spike activity in cerebral cortex. IEEE Trans Neural Syst Rehabil Eng. 2003;11:151–155. doi: 10.1109/TNSRE.2003.814443. [DOI] [PubMed] [Google Scholar]
  • 19.Li Y, Baek K, Gulari M, Wise K. A drug-delivery probe with an in-line flowmeter based on trench refill and chemical mechanical polishing techniques. Sensors, 2007 IEEE. 2007:1144–1147. [Google Scholar]
  • 20.Lieberman DM, Laske DW, Morrison PF, Bankiewicz KS, Oldfield EH. Convection-enhanced distribution of large molecules in gray matter during interstitial drug infusion. J Neurosurg. 1995;82:1021–1029. doi: 10.3171/jns.1995.82.6.1021. [DOI] [PubMed] [Google Scholar]
  • 21.Linninger AA, Somayaji MR, Mekarski M, Zhang L. Prediction of convection-enhanced drug delivery to the human brain. J Theor Biol. 2008;250:125–138. doi: 10.1016/j.jtbi.2007.09.009. [DOI] [PubMed] [Google Scholar]
  • 22.Ludwig KA, Miriani RM, Langhals NB, Joseph MD, Anderson DJ, Kipke DR. Using a common average reference to improve cortical neuron recordings from microelectrode arrays. J Neurophysiol. 2009;101:1679–1689. doi: 10.1152/jn.90989.2008. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Ludwig KA, Uram JD, Yang J, Martin DC, Kipke DR. Chronic neural recordings using silicon microelectrode arrays electrochemically deposited with a poly(3,4-ethylenedioxythiophene) (PEDOT) film. J Neural Eng. 2006;3:59–70. doi: 10.1088/1741-2560/3/1/007. [DOI] [PubMed] [Google Scholar]
  • 24.Maggi C, Meli A. Suitability of urethane anesthesia for physiopharmacological investigations in various systems part 1: General considerations. Cellular and Molecular Life Sciences (CMLS) 1986;42:109–114. doi: 10.1007/BF01952426. [DOI] [PubMed] [Google Scholar]
  • 25.Maggi C, Meli A. Suitability of urethane anesthesia for physiopharmacological investigations in various systems. part 2: Cardiovascular system. Cellular and Molecular Life Sciences (CMLS) 1986;42:292–297. doi: 10.1007/BF01942510. [DOI] [PubMed] [Google Scholar]
  • 26.Martin JH, Ghez C. Pharmacological inactivation in the analysis of the central control of movement. J Neurosci Methods. 1999;86:145–159. doi: 10.1016/s0165-0270(98)00163-0. [DOI] [PubMed] [Google Scholar]
  • 27.Morrison PF, Chen MY, Chadwick RS, Lonser RR, Oldfield EH. Focal delivery during direct infusion to brain: Role of flow rate, catheter diameter, and tissue mechanics. Am J Physiol. 1999;277:R1218–29. doi: 10.1152/ajpregu.1999.277.4.R1218. [DOI] [PubMed] [Google Scholar]
  • 28.Narahashi T, Moore JW, Scott WR. Tetrodotoxin blockage of sodium conductance increase in lobster giant axons. J Gen Physiol. 1964;47:965–974. doi: 10.1085/jgp.47.5.965. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Narahashi T. Mechanism of action of tetrodotoxin and saxitoxin on excitable membranes. Fed Proc. 1972;31:1124–1132. [PubMed] [Google Scholar]
  • 30.Neeves KB, Lo CT, Foley CP, Saltzman WM, Olbricht WL. Fabrication and characterization of microfluidic probes for convection enhanced drug delivery. J Control Release. 2006;111:252–262. doi: 10.1016/j.jconrel.2005.11.018. [DOI] [PubMed] [Google Scholar]
  • 31.Nicholson C. Diffusion and related transport mechanisms in brain tissue. Rep Prog Phys. 2001;64:815–884. [Google Scholar]
  • 32.Papageorgiou D, Shore S, Bledsoe S, Jr, Wise K. A shuttered neural probe with on-chip flowmeters for chronic in vivo drug delivery. Microelectromechanical Systems, Journal of. 2006;15:1025–1033. [Google Scholar]
  • 33.Polikov VS, Tresco PA, Reichert WM. Response of brain tissue to chronically implanted neural electrodes. J Neurosci Methods. 2005;148:1–18. doi: 10.1016/j.jneumeth.2005.08.015. [DOI] [PubMed] [Google Scholar]
  • 34.Popelar J, Nwabueze-Ogbo FC, Syka J. Changes in neuronal activity of the inferior colliculus in rat after temporal inactivation of the auditory cortex. Physiol Res. 2003;52:615–628. [PubMed] [Google Scholar]
  • 35.Rathnasingham R, Kipke DR, Bledsoe SC, Jr, McLaren JD. Characterization of implantable microfabricated fluid delivery devices. IEEE Trans Biomed Eng. 2004;51:138–145. doi: 10.1109/TBME.2003.820311. [DOI] [PubMed] [Google Scholar]
  • 36.Retterer S, Smith K, Bjornsson C, Turner J, Isaacson M, Shain W. Constant pressure fluid infusion into rat neocortex from implantable microfluidic devices. Journal of Neural Engineering. 2008;5:385–391. doi: 10.1088/1741-2560/5/4/003. [DOI] [PubMed] [Google Scholar]
  • 37.Retterer ST, Smith KL, Bjornsson CS, Neeves KB, Spence AJ, Turner JN, et al. Model neural prostheses with integrated microfluidics: A potential intervention strategy for controlling reactive cell and tissue responses. IEEE Trans Biomed Eng. 2004;51:2063–2073. doi: 10.1109/TBME.2004.834288. [DOI] [PubMed] [Google Scholar]
  • 38.Spataro L, Dilgen J, Retterer S, Spence AJ, Isaacson M, Turner JN, et al. Dexamethasone treatment reduces astroglia responses to inserted neuroprosthetic devices in rat neocortex. Exp Neurol. 2005;194:289–300. doi: 10.1016/j.expneurol.2004.08.037. [DOI] [PubMed] [Google Scholar]
  • 39.Sykova E, Nicholson C. Diffusion in brain extracellular space. Physiol Rev. 2008;88:1277. doi: 10.1152/physrev.00027.2007. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.van Duuren E, van der Plasse G, van der Blom R, Joosten RN, Mulder AB, Pennartz CM, et al. Pharmacological manipulation of neuronal ensemble activity by reverse microdialysis in freely moving rats: A comparative study of the effects of tetrodotoxin, lidocaine, and muscimol. J Pharmacol Exp Ther. 2007;323:61–69. doi: 10.1124/jpet.107.124784. [DOI] [PubMed] [Google Scholar]
  • 41.Velliste M, Perel S, Spalding MC, Whitford AS, Schwartz AB. Cortical control of a prosthetic arm for self-feeding. 2008 doi: 10.1038/nature06996. [DOI] [PubMed] [Google Scholar]
  • 42.Zhuravin IA, Bures J. Extent of the tetrodotoxin induced blockade examined by pupillary paralysis elicited by intracerebral injection of the drug. Exp Brain Res. 1991;83:687–690. doi: 10.1007/BF00229849. [DOI] [PubMed] [Google Scholar]

RESOURCES