Characterization of Motor and Somatosensory Evoked Potentials in the Yucatan Micropig Using Transcranial and Epidural Stimulation

Abstract

Yucatan micropigs have brain and spinal cord dimensions similar to humans and are useful for certain spinal cord injury (SCI) translational studies. Micropigs are readily trained in behavioral tasks, allowing consistent testing of locomotor loss and recovery. However, there has been little description of their motor and sensory pathway neurophysiology. We established methods to assess motor and sensory cortical evoked potentials in the anesthetized, uninjured state. We also evaluated epidurally evoked motor and sensory stimuli from the T6 and T9 levels, spanning the intended contusion injury epicenter. Response detection frequency, mean latency and amplitude values, and variability of evoked potentials were determined. Somatosensory evoked potentials were reliable and best detected during stimulation of peripheral nerve and epidural stimulation by referencing the lateral cortex to midline Fz. The most reliable hindlimb motor evoked potential (MEP) occurred in tibialis anterior. We found MEPs in forelimb muscles in response to thoracic epidural stimulation likely generated from propriospinal pathways. Cranially stimulated MEPs were easier to evoke in the upper limbs than in the hindlimbs. Autopsy studies revealed substantial variations in cortical morphology between animals. This electrophysiological study establishes that neurophysiological measures can be reliably obtained in micropigs in a time frame compatible with other experimental procedures, such as SCI and transplantation. It underscores the need to better understand the motor control pathways, including the corticospinal tract, to determine which therapeutics are suitable for testing in the pig model.

Introduction

Evoked potential (EP) measurements have been used to assess residual connectivity after spinal cord injury (SCI) in people and can correlate with the probability of neurological recovery.^1,2 They are included in the National Institute of Neurological Disorders and Stroke common data elements for SCI.³ Somatosensory and motor evoked potentials (SSEPs/MEPs) are used in operative neurosurgery to avoid damage to motor and sensory pathways.^4,5 EPs allow real-time detection of axonal conduction changes caused by altered spinal cord signal transmission as may occur during ischemia or during tumor resections and deformity surgery.^5
–7 EPs are recognized as among the only early biomarkers to predict postoperative status after spinal cord interventions.⁸

Yucatan micropigs are emerging as a useful model for SCI studies.⁹ Though their brains have been mapped for stereotaxic purposes, little is known regarding the anatomy and neurophysiology of their spinal cord tracts.^10
–12 This information is important to advance the model. For human SCI therapeutics research, EPs may be valuable tools to assess the tolerance of injured spinal cord pathways to treatments, such as intraparenchymal cell transplants. For these interventions, residual signals post-SCI can provide a baseline to determine whether an intervention compromises retained connectivity. An example is whether the volume and rate of delivery of a cellular injection exceeds a critical threshold and causes additional injury, evidenced by sustained reduced conduction. Such methods are infrequently used in SCI clinical trials, exposing patients to increased risk to their preserved neurological function. This type of safety testing has been studied for intracerebral injections in the Gottingen minipig model.¹³

EPs require both time and technical proficiency. Intraoperative testing should be conducted in a time frame compatible with other study procedures, not excessively prolonging anesthesia. Reliable, accurate, and rapid testing requires experience and development of suitable methodology. We sought to develop robust methods for EP stimulation and recording in the Yucatan model.^9,14,15 In addition, we assessed the time required to conduct different combinations of the tests because both the risk for complications, and costs, are increased with the duration of surgical procedures.^16,17

During this study, aspects of porcine MEPs and SSEPs, such as pathways, detection frequency, reproducibility, latency, and amplitude, were characterized across the sample. We also studied the motor and sensory EPs elicited by epidural stimulation of the dorsal surface of the thoracic spinal cord at sites selected to span a subsequent contusion injury. We hypothesized that epidural stimulation (ES) would be associated with larger amplitudes and superior signal-to-noise ratios increasing the sensitivity of EP detection. The muscle detection sensitivity of transcranial versus epidural evoked potentials was determined. Because our current porcine model is based on thoracic SCI and treatments to increase locomotor recovery, this study emphasized the detection of lower extremity muscles. We focused on an electromyography (EMG) readout, as compared to direct spinal cord potentials, with a view to repeated longitudinal assessments, not requiring subsequent surgical exposure of the spinal cord.

Methods

Experimental design

The experiments assessed the feasibility to acquire reliable EPs before and after laminectomy and before thoracic contusive SCI. To define reference values, we determined the consistency of EPs between animals that differed, to some extent, in age and weight. EPs were obtained from 26 animals in four ways: 1) transcranial electrical stimulation (TcES) to evoke MEPs in limb muscles; 2) transcranial somatosensory recordings (SSEPs) elicited by repetitive peripheral nerve stimulation; 3) MEPs; and 4) SSEPs evoked by epidural stimulation (Fig. 1). Of 26 animals, 10 underwent all four testing modalities. In 8, we focused on testing TcES compared to epidural motor stimulation (motor group), and in 8 epidurally and peripheral nerve evoked SSEPs were compared (sensory group).

FIG. 1.

Key landmarks for the porcine electrophysiology protocol (A and B). Photographs from cadaveric dissections, showing the porcine brain and its relation to external landmarks. The sagittal midline was measured and traced on the head. The nasion was established as a line extending between the medial palpebral commissures. The inion was marked over the most posterior prominence of the occipital bone, the distance between nasion-inion lines was measured, and the midpoint marked as the vertex. Screws were placed 0.5–1.0 cm caudal and 0.5–1.0 cm cranial from the vertex and 0.5–1.0 cm from the midline. (B) Porcine cortex; the black arrowhead points to the cruciate gyrus, and the two arrows are pointing at coronal (white) and cruciate sulci (black). Note that the cruciate gyrus continues with the coronal gyrus without a sulcal division. (C) Electrode screws placed in an anesthetized subject, in proximity to underlying motor and sensory cortices. (D and E) Post-mortem exposure of hindlimb muscles. The two positions used for comparison of the TA EMG recordings are shown (dots lateral position, asterisk midline position). In €, arrows show the intended position for the recording needles to monitor AH EMG. (F and G) Schemata of EP modalities used in the protocol showing recording and stimulation points. AH, abductor hallucis; EMG, electromyography; EP, evoked potential; GAS, gastrocnemius; LDE, lateral digital extensor; MEPs, motor evoked potentials; SSEPs, somatosensory evoked potentials; TA, tibialis anterior; TcES, transcranial electrical stimulation.

The more limited testing helped determine the time-saving as compared to testing all four modalities. However, each animal had at least a minimum baseline motor and sensory data set. For each type of EP, the percentage of successful recordings/total stimuli and the range of values were determined. The relative sensitivity and consistency of the EP tests was ascertained. After the baseline assessments, animals received a spinal cord contusion. Injury-related conduction changes will be discussed in another report. Preliminary data related to intracortical and direct wave stimulation are provided to address issues related to the corticospinal tract.

Procedures with animals

The experimental procedures were approved by the Institutional Animal Care and Use Committee of the University of Miami (Miami, FL). Twenty-six female Micro-Yucatan miniature Swine (Sinclair BioResources, Windham, ME), ages 5–10 months, weighing between 15 and 25 kg, were housed, fed, and cared for in accord with United States Department of Agriculture and Association for Assessment and Accreditation of Laboratory Animal Care regulations. Female animals were selected because their urethra and bladders are much easier to manage than males post-SCI. Anesthesia was induced with ketamine/xylazine (30–45 mg/kg, 2–4 mg/kg). Pigs were intubated, ventilated, and maintained with low doses of isoflurane (1–2%) during placement of intra-arterial and venous lines. A 3-lead EKG was collected, intra-arterial blood pressure was monitored (using a sensor placed inside the superficial branch of the femoral artery), and a central line used to deliver fluids and drugs. Cefazolin 30 mg/kg intravenously was administered for antibiotic prophylaxis and post-operatively every 8 h for up to 24 h. Temperature was regulated using Bair Hugger blankets in response to rectal temperature measurements. For EP testing, gas anesthesia was switched to a total intravenous anesthesia (TIVA) protocol that included: propofol (25–150 mcg/kg/min); ketamine (10–180 mcg/kg/min); and fentanyl (25–100 mcg/kg/h). TIVA was continuously regulated to ensure adequate anesthetic depth and yet enable effective stimulation and recording.

Intraoperative electrophysiology

Cadaveric studies were performed to correlate key intra- and extracranial landmarks. Previous studies suggested that the motor cortex in the pig is in the cruciate gyrus between coronal and cruciate sulci,^12,18,19 and the limb somatosensory cortex extends onto the interhemispheric face of the cruciate sulcus (Fig. 1A).²⁰ Reference external skull marks were traced (e.g., nasion, inion, vertex, and midline) to establish where to place electrode screws over the desired cortex. We adapted the international 10–20 nomenclature; FC1 and FC2 were assigned to points 7.5 mm toward the nasion from vertex and 5 mm lateral to the midline for motor stimulation. For sensory recordings, CP1 and CP2 were assigned to points 7.5 mm toward the inion 5 mm lateral to the midline. Fz was assigned to a midline reference point 2 cm anterior to the vertex and used to improve the lateralization of SSEP detection.

To reduce impedance and improve both motor cortex stimulation and signal recordings from the somatosensory cortex, two small incisions of 1 cm were made in the skin over the described positions. Superficial pilot holes were drilled in the skull and sterilized stainless steel screws (15 mm length, 1.1 mm diameter) with a flattened tip were advanced to serve as electrodes. To avoid complications such as infection, the electrode screws were removed at the end of EP recording and the skin was approximated with surgical tape. Three animals at the beginning of the study were assessed using human intraoperative surface corkscrew needles in contact with the skull surface. Only small and inconsistent SSEPs were recorded with high voltages needed to obtain MEPs. These experiments are not included in this report because placing cranial screws through the bone led to much better recordings.

Somatosensory evoked potentials from peripheral nerve stimulation

Stimulation was applied to the tibial and median nerves of each side by placing subdermal needle electrodes (1.2 cm, 27 Gauge, Neuroline twisted pair; Ambu, Copenhagen, Denmark) near the usual location of the nerves. Stimulation sites were refined by trial-and-error testing, assessing twitch strength in the muscles of the hooves. Hindlimbs and forelimbs were stimulated separately in a left-to-right interleaved sequence and recorded in the contralateral cortex. Repetitive stimuli (1.1 Hz, 300 μs) were given at intensities twice the threshold required to visualize twitching in the muscles distal to the stimulation point. For somatosensory recordings, skull screws were connected to the recording channels in referential mode. The signal was filtered (10-Hz low frequency, 300-Hz high frequency, Protector IOM, Xltek; Natus Medical Inc., Pleasanton, CA) with an average of 1000 sweeps recorded. Recording arrays used were: CP2–CP1 and CP2–Fz for the left tibial nerve and CP1–CP2 and CP1–Fz for the right tibial nerve. Sciatic nerve potentials were acquired to verify adequate tibial nerve stimulation. Long needle electrodes with a 3-mm uninsulated tip (7.5 cm, 25 gauge, Neuroline monopolar; Ambu) were placed through the sciatic notch in each side. Median nerve stimulation was performed in a similar manner using the same cortical recording arrangements. The stimulation and recording parameters selected resemble those used in human neuromonitoring.²¹

Transcranial motor evoked potentials

To depolarize the motor cortex, stimulation trains were delivered through a stainless steel alligator clip clamped to the electrode screws (Fig. 1C) with an impedance <2.5 kOhms. Stimuli intensities ranged from 200 to 400 V in trains of five pulses (pulse duration, 0.5 ms; interstimulation interval [ISI], 5.0 ms; Protector IOM, Xltek, internal stimulator; Natus Medical Inc.). For hindlimb muscle depolarization, stimulation parameters were assessed using the anode over the right cortex and the cathode over the left side (FC2–FC1). This stimulus arrangement elicited bilateral depolarization. Multi-pulse stimulation was necessary to overcome the effects of the anesthetics and increase motorneuron recruitment.^6,22
–24

Free running EMG was observed during MEP collection to ensure minimal background noise. The filters for EMG were the same as for MEPs, but set to epochs of 10 sec. For muscle evoked recordings and EMG, two types of needle electrodes were used: 1) small bipolar needles for superficial small muscles (1.2 cm, 27 gauge, Neuroline twisted pair; Ambu) and 2) a combination of two longer monopolar needles for larger deep muscles (5 cm, 25 gauge, Neuroline; Ambu). Needle electrodes were placed in the muscles of the left and right forelimbs (extensor carpi radialis; ECR) and hindlimbs (gastrocnemius [GAS], tibialis anterior [TA], biceps femoris [BF], vastus lateralis [VL], and rectus femoris [RF]; Fig. 1D,E). Additional hindlimb muscles were tested in some pigs during MEP recordings: lateral digital extensor (LDE; (n = 6); vastus medialis (n = 8); and abductor hallucis (AH) in 10 pigs. In 16 animals, the effect of positioning needle electrodes in the lateral belly versus the midline of the TA muscle was compared (Fig. 1D) to assess the importance of being in proximity to the exact motor point. Signal recorded from the muscles was amplified and then filtered using settings suggested in previous reports (1.5-Hz low frequency, 875-Hz high frequency, Protector IOM, Xltek; Natus Medical Inc.).^22,23

Spinal cord exposure and epidural lead placement

The skin was sterilely prepped with chlorhexidine and an incision created between T5 and T10. A consistent surface counting method was used to identify the laminectomy level, verified in some animals using fluoroscopy, and after perfusion, by counting nerve roots. Muscles were detached from spinous processes, retractors placed, and a complete T7/T8 laminectomy created to expose spinal cord segments T8/9. With the dura exposed, two double-contact strip electrodes (custom made, width 7.3 mm, length 4 cm with contact diameters 4.5 mm and rated impedance <150 ohms; Cortac Electrodes; PMT Corporation, Chanhassen, MN) were slipped into the dorsal epidural space 1 cm cranial and 1 cm caudal to the laminectomy boundaries. The remaining ligamentum flavum and epidural fat maintained the electrodes in contact with the dura mater. These levels were selected to span the planned spinal cord contusion injury. For this report, we consider these electrodes to stimulate over T6 and T9, but they could have partially reached into the rostral T5 and distal T10 segments.

Motor and sensory epidural EPs were triggered and recorded. Next, the contusive injury was created. Before closing, the electrodes were removed and the incision closed in layers. In 1 control animal, a sequential transection was performed while collecting EPs to assess for possible false-positive conduction after complete transection. In this pig, we acutely determined the minimum substrate of residual spinal cord that could convey signals. For transection, the dura was incised and tacked laterally to allow cord visualization. Sequential precision incisions were made into the cord, first in the left half and then in the remaining right half. Under the high magnification provided by a surgical microscope, it was possible to distinguish the midline and gray and white matter. After transection and EP recording, the dura was repaired and the incision closed.

Somatosensory evoked potentials from epidural stimulation

SSEPs were triggered from the spinal cord electrodes using stimuli of 5–10 mA, 1.3 Hz, and 200-μs intervals. Recordings averaged 1000 sweeps, using the same cortical recording sites as described for peripheral nerve stimulation. ES of the dorsal spinal cord stimulates both sides; thus, bilateral CP2–CP1 and unilateral CP2–Fz, CP1–Fz were used to compare activity within the left and right sensory cortices (being CP1–CP2 redundant and therefore not recorded). Several groups have reported stimulation over the dorsal columns to elicit SSEPs^25

–29; nonetheless, because the parameters varied between reports, stimulation/recording parameters presented here were adapted from peripheral nerve SSEP techniques.

Epidural motor evoked potentials

Epidural motor stimulation (ES) was performed independently at T6 and T9 using reported parameters,³⁰ trains of three or five pulses (0.05 ms, ISI 0.5 ms) were delivered at intensity ranges of 30–50 V (MultiPulse Stimulator D-185; Digitimer Ltd., Hertfordshire, UK). Signals were recorded from the same muscles used for transcranial MEP (TcMEP) testing to obtain comparable MEPs.

Preliminary assessments

Later in the study we conducted three preliminary assessments: 1) Attempts to record D waves from the thoracic dural surface. Detecting D waves indicates a corticospinal tract at the recorded level in the spinal cord. 2) Direct intracortical stimulation to assess the forelimb and hindlimb locations for the injection of anterograde tracers to define the corticospinal tract extent. 3) Attempts to measure conduction between the two epidural electrodes by stimulating one and recording from the other. This might be a useful post-injury technique to test for retained conduction through the injury epicenter.

Statistical analysis

Somatosensory evoked potentials

Typical values were established from cortical recordings obtained in 18 animals in which tibial, median, and epidural stimuli were applied. The recording arrays CP1–CP2 and CP2–CP1 and CP1–Fz and CP2–Fz were compared for signal resolution quality. Positive and negative peaks of SSEPs were marked and their latency value in milliseconds was determined. Averages and standard deviations were calculated for the consistent peaks, P18 and N24 for median nerve and epidural stimulation, and P29 and N35 for tibial nerve stimulation. Amplitudes were measured in segments P14–N24 and P29–N35. P14 values out of the 10- to 30-ms range and peak-to-peak amplitude values <0.2 uV were excluded.

Motor evoked potentials

Amplitudes were measured using peak-to-peak values and latencies marked at the beginning of the first deflection. The presence, latency, and amplitude of the most similar four of six collected waveforms from each monitored muscle were calculated and compared among different animals to establish the typical range for each muscle. Samples with amplitudes <40 μV and latencies were <6 ms were excluded from analysis. To determine which muscles were more responsive to cortical stimulation under the experimental conditions, the percentage of detection from each side was determined.

Epidural MEPs (EpiMEPs) obtained while stimulating the cord above and below the laminectomy were measured in 18 animals. The detection rate of TcMEPs and EpiMEPs and their amplitudes and latencies were compared to ascertain which form of motor stimulation most reliably triggered hindlimb muscles. The mean difference between the latencies of MEP onset for each muscle from T6 and T9 was calculated. These values will allow subsequent comparison of MEPs post-SCI.

To determine whether the differences between recorded values were sufficiently consistent to be considered independent, we applied statistical tests. The t-test and one-way analysis of variance (ANOVA) were used to compare two or more groups respectively, with p values ≤0.05 considered significant. Excel (2010; Microsoft, Inc., Redmond, WA) and GraphPad software (version 4.1; GraphPad Software Inc., La Jolla, CA) were used for analysis and graphical representation of the data.

Results

In this study, no animals experienced obvious seizures. In a previous animal, not part of this study, 1 animal had a surgical complication related to neuromonitoring and cortical stimulation. In this animal, a probable brain abscess was detected on post-mortem exam. This problem had been clinically silent, but appeared to be linked to the placement of a skull screw. In this animal, the screws were attached permanently. After detecting this complication, we removed the screws at the close of each procedure. The surgical exposure of the spinal cord did not abolish ascending or descending evoked potentials in any animal, verifying the safety of the laminectomy procedure. This is important so that subsequently observed changes can be interpreted to be caused by the contusion event. We emphasize the average findings across the 26-animal sample because the utility of the testing depends on its reproducibility in the population, not just in individual animals. Latencies showed greater similarity between animals than amplitudes.

Somatosensory evoked potential findings

SSEP assessments were reliable (summarized in Table 1). In Figure 2A–C, the characteristic features of SSEPs obtained using midline spanning and midline referenced (Fz) cortical arrays in response to peripheral nerve and epidural stimulation is shown. Control values recorded from the sciatic (or popliteal) nerve verified effective tibial nerve stimulation. Detection percentages were calculated as the incidence of clearly definable peaks detected over the total number of animals tested (18). The latency value is shown for each peak and the segment P1–N1. Tibial and sciatic nerve conduction latencies to the cortex were used to estimate the central sensory conduction velocity (CSCV). An average value of 46.4 m/s was estimated by applying the formula CSCV = total distance (tibial nerve to cortex / tibial nerve-evoked SSEP latency – sciatic nerve latency).

FIG. 2.

Characteristics of porcine evoked potentials. (A) Median nerve SSEPs; two different cortical arrays are presented per side (e.g., CP1–CP2, CP1–Fz), and N24 has been marked as a component recorded from the cortex contralateral to the stimulation side. In this example, an extra point was monitored at the fifth cervical level, “CS5,” and referenced to Fz. The resultant far-field recording is interpreted as extracortical and indicates the cervicomedullary origin of the first negative peak (P14) in the cortical recordings. (B) Epidural SSEPs; T6 stimulation generated an N24 peak, whereas T9 generated a smaller amplitude signal with a 2-ms greater latency. The Fz reference point gives better amplitude resolution (e.g., CP2–CP1 vs CP2–FZ. (C) Tibial nerve SSEPs, showing cortical recordings with an N28 (27 ms) in the contralateral referenced array and N35 (32 ms) in the Fz referenced array. Note the smaller amplitude as compared to median nerve SSEPs. An early N5 was recorded from the popliteal nerve, arrowheads. (D) MEPs collected in the TA muscle bilaterally during TcMEP and epidural stimulation; note the shorter latency recorded as the result of lower segment stimulation compared to TcMEP. (E) Bilateral TA MEPs showing the different morphology obtained by locating two sets of needle electrodes within the same muscle. Latencies and amplitudes remain unchanged. Epi, epidural; LTA, left tibialis anterior; MEPs, motor evoked potentials; RTA, right tibialis anterior; SSEPs, somatosensory evoked potentials; TA, tibialis anterior; TcMEP, transcranial motor evoked potential.

Table 1.

SSEP Cortical Array Reliability and Nerve and Neurogenic Stimulation in the Micro-Yucatan Miniature Swine

			Tibial nerve				Median nerve				T6 Stimulation				T9 Stimulation
Array	Parameter measured	Peak	%	Mean	SD	SEM	%	Mean	SD	SEM	%	Mean	SD	SEM	%	Mean	SD	SEM
CP2–Fz	Peaks Lat (ms)	P1	94.4	28.4	3.4	0.8	100	18.6	1.8	0.4	100	16.0	2.4	0.6	100	17.3	3.3	0.8
		N1	100	34.8	2.9	0.7	100	24.9	2.4	0.6	100	22.9	2.7	0.6	100	24.3	3.7	0.9
		P2	100	44.1	3.2	0.8	100	34.1	4.7	1.1	100	31.7	4.0	0.9	100	32.6	9.4	2.2
	Segment Lat (ms)	P1–N1		6.3	1.5	0.4		6.3	1.4	0.4		6.9	1.7	0.4		7.0	1.5	0.4
	Peak segmentAmp (uV)	P1–N1		0.8	0.5	0.1		7.2	5.0	1.2		7.2	7.4	1.8		4.7	3.4	0.8
		N1–P2		0.9	0.5	0.1		6.5	4.7	1.1		7.8	6.6	1.6		5.4	4.1	1.0
CP1–Fz	Peaks Lat (ms)	P1	88.9	27.7	5.0	1.3	100	19.9	2.2	0.5	100	15.9	2.6	0.6	100	17.2	3.5	0.8
		N1	100	35.2	4.6	1.1	100	26.8	5.3	1.2	100	22.7	2.4	0.6	100	23.5	4.3	1.0
		P2	100	43.9	5.0	1.2	94.4	34.1	3.4	0.8	100	31.5	2.92	0.7	100	33.6	5.4	1.3
	Segment Lat (ms)	P1–N1		7.6	2.7	0.7		5.7	1.2	0.3		7.0	1.3	0.3		6.3	2.4	0.6
	Peak segmentAmp (uV)	P1–N1		0.9	0.6	0.2		5.3	4.3	1.00		7.3	6.2	1.5		5.7	5.3	1.3
		N1–P2		0.1	0.6	0.2		4.0	2.7	0.7		8.6	6.2	1.5		7.1	7.5	1.8
CP2–CP1	Peaks Lat (ms)	P1	100	27.8	3.4	0.8	100	18.7	1.6	0.4	100	16.9	3.1	0.7	100	17.6	3.5	0.8
		N1	100	34.9	2.7	0.6	100	24.4	2.4	0.6	100	23.6	4.1	1.0	100	24.0	4.6	1.1
		P2	94.4	42.9	3.2	0.8	100	34.4	3.9	0.9	100	32.2	5.4	1.3	100	32.5	4.2	1.0
	Segment Lat (ms)	P1 - N1		7.1	2.4	0.6		5.7	1.3	0.3		6.8	2.2	0.5		6.4	2.1	0.5
	Segment Amp (uV)	P1–N1		0.8	0.3	0.1		5.0	4.1	1.0		3.1	2.6	0.6		2.3	1.5	0.4
		N1–P2		1.0	0.6	0.1		5.0	3.9	0.9		3.1	1.6	0.4		2.4	1.3	0.3
CP1–CP2	Peaks Lat (ms)	P1	88.9	27.3	4.6	1.1	94.4	19.0	1.7	0.4
		N1	100	34.0	3.6	0.9	100	25.9	2.9	0.7
		P2	100	44.0	4.9	1.2	100	35.3	4.8	1.1
	Segment Lat (ms)	P1–N1		7.1	2.9	0.7		6.3	1.4	0.4
	Peak segmentAmp (uV)	P1–N1		0.7	0.4	0.1		3.5	2.2	0.5
		N1–P2		1.1	0.6	0.1		3.2	2.4	0.6
Right sciatic nerve	Peaks Lat (ms)	P1	83.3	4.1	0.6	0.2
		N1	83.3	4.8	0.6	0.2
		P2	83.3	5.6	0.7	0.2
	Peak segmentAmp (uV)	P1–N1		7.6	8.8	2.4
		N1–P2		9.3	10.3	2.8
Left sciatic nerve	Peaks Lat (ms)	P1	88.9	4.0	0.3	0.1
		N1	88.9	4.7	0.3	0.1
		P2	83.3	5.3	0.4	0.1
	Peak segmentAmp (uV)	P1–N1		12.0	8.6	2.2
		N1–P2		12.7	11.6	3.0

SSEP, somatosensory evoked potential; Lat, latency; Amp, amplitude; SD, standard deviation; SEM, standard error of the mean.

Comparing the latencies of positive and negative somatosensory evoked potential peaks

In all stimulation/recording modalities, a positive peak designated as P1 had the shortest latency, followed by N1 and P2. The average latencies from median and tibial nerves were P1 (19 and 27 ms), N1 (24 and 34 m), and P2 (34 and 43 ms). P1, N1, and P2 latencies differed statistically (one-way ANOVA; p < 0.0001), substantiating their independence. Latency values of SSEP peaks (P1, N1, and P2) were consistent between animals in response to epidural, tibial, and median nerve stimulation. Thoracic ES generated responses with latencies similar to the median nerve. There was no difference in latency consistency between cross-midline or unilaterally referenced skull electrode arrays. The low relative standard error of the mean (SEM) percent (SEM/latency) for the SSEPs of 1–3% indicates that the mean values are useful references for the population.

Amplitudes recorded from Fz referenced versus contraipsilateral arrays

SSEP segment amplitudes P1–N1 and N1–P2 were larger in midline Fz versus cross-midline arrays for both T6 and T9 ES (t-test; p < 0.001). However, for the tibial and median nerves, there was no significant difference in either amplitude or latency values between these arrays. Amplitudes from tibial nerve stimulation were 8 × smaller ( \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document}$$\overline {x}$$ \end{document} = 1.01 ± 0.17 μV standard deviation [SD]) than epidurally stimulated SSEPs (e.g., CP1–Fz recording in response to T6 stimulation; \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document}$$\overline {x}$$ \end{document} = 8.6 ± 6.2 μV SD). Median nerve and epidural stimulation had similar amplitude values. The largest response to median nerve simulation was recorded with the CP2–Fz array ( \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document}$$\overline {x}$$ \end{document} = 7.2 ± 4.5 μV SD). Recordings >2 μV in amplitude were easier to identify because a lower noise/signal ratio.

Percentage detection of somatosensory evoked potentials

ES was very effective to generate high-amplitude SSEPs from both T6 and T9 with consistent and definable peaks in all recording arrays and animals. Regarding waveform, the second peak N was more definite than the first peak P. In 6% of animals, the first peak was not detected from median nerve stimulation when recording with the CP1–Fz array. All other arrays were positive for both P and N in 100% of the animals.

Motor evoked potential findings

MEPs require combinations of events, neuronal-axonal excitation, and motor neuron firing and are more sensitive to inhalational anesthesia than SSEPs.³¹ We determined the most consistently detected of nine muscles in response to either cranial or epidural stimulation. In Table 2, muscles are ranked based on MEP detection frequency. Means, SD, and SEM are shown. The percentage detection of both muscles (e.g., 18 animals, 36 muscles) is indicated by M, whereas the percent detection of at least one of the two muscles is indicated by S. In Figure 2, panels D and E show the usual shape and consistency of MEPs recorded from TA in response to transcranial stimulation and ES at T6 and T9.

Table 2.

MEP Measurement Reliability Per Muscle in the Micro-Yucatan Miniature Swine Using Transcranial and Neurogenic Stimulation

		TcMEP measurements				Epidural MEP (T6)				Epidural MEP (T9)
Muscles	Measure	Detection % (n)	Mean	SD	SEM	Detection % (n)	Mean	SD	SEM	Detection % (n)	Mean	SD	SEM
Tibialis anterior	Amp (uV)	M 100 (36)	549.4	281.6	66.4	M 97.2 (36)	1036.9	742.3	174.9	M 100 (36)	1138.3	670.6	158.1
	Lat (ms)	S 100 (18)	27.1	4.6	1.1	S 100 (18)	20.5	3.9	0.9	S 100 (18)	17.9	3.4	0.8
Biceps femoris	Amp (uV)	M 91.6 (36)	536.3	576.7	139.9	M 86.1 (36)	552.3	413.2	100.2	M 86.1 (36)	505.0	381.5	92.5
	Lat (ms)	S 94.4 (18)	26.3	4.7	1.1	S 94.4 (18)	18.9	5.6	1.4	S 94.4 (18)	16.8	4.5	1.1
Lateral digital extensor	Amp (uV)	M 75.0 (12)	268.7	338.4	151.4	M 75 (12)	256.8	261.5	116.9	M 91.7 (12)	467.8	469.1	191.5
	Lat (ms)	S 83.3 (6)	27.9	5.6	2.5	S 83.3 (6)	22.5	5.2	2.3	S 100 (6)	20.1	4.1	1.7
Gastrocnemius	Amp (uV)	M 72.2 (36)	291.9	332.8	83.2	M 83.3 (36)	416.3	388.5	94.2	M 80.5 (36)	413.1	380.2	92.2
	Lat (ms)	S 88.8 (18)	27.6	4.6	1.2	S 94.4 (18)	20.9	4.4	1.1	S 94.4 (18)	18.3	3.2	0.8
Abductor hallucis	Amp (uV)	M 45 (20)	369.1	510.5	193.0	M 75 (20)	900.9	720.6	227.9	M 80 (20)	920.3	622.4	207.5
	Lat (ms)	S 60 (10)	32.2	4.9	1.9	S 100 (10)	23.6	5.8	1.8	S 90 (10)	20.1	3.6	1.2
Vastus medialis	Amp (uV)	M 25.0 (16)	1114.0	190.5	134.5	M 68.8 (16)	279.3	311.4	127.1	M 62.5 (16)	578.7	452.4	171.0
	Lat (ms)	S 25.0 (8)	23. 6	8.5	5.9	S 75 (8)	17. 5	6.4	2.6	S 87.5 (8)	14.6	4.2	1.6
Vastus lateralis	Amp (uV)	M 19.4 (36)	564.9	380.5	190.2	M 77.8 (36)	272.2	292.8	75.6	M 75 (36)	406.9	480.5	124.1
	Lat (ms)	S 27.7 (18)	27.3	8.1	3.6	S 83.3 (18)	16.4	5.2	1.4	S 83.3 (18)	14.6	5.8	1.5
Rectus femoris	Amp (uV)	M 13.8 (36)	522.9	486.8	234.4	M 55.6 (36)	356.4	381.2	105.7	M 72.2 (36)	404.0	373.8	99.9
	Lat (ms)	S 22.2 (18)	19.5	5.8	2.9	S 72.2 (18)	15.6	4.9	1.4	S 77.8 (18)	13.5	4.4	1.2
Extensor carpi radialis	Amp (uV)	M 100 (36)	1655.4	689.2	162.4	M 88.9 (36)^a	1005.4	547.5	136.9	M 50 (36)^a	541.1	388.9	103.9
	Lat (ms)	S 100 (18)	15.8	5.2	1.2	S 88.9 (18)^a	12.6	3.1	0.8	S 55.6 (18)^a	17.4	4.6	1.4

MEP measurement reliability per muscle in the Micro-Yucatan miniature Swine using transcranial and neurogenic stimulation. N = total number of animals in which a muscle was tested; M = positive signal per muscle (left or right); S = positive signal per subject (bilateral).

Only early response values were calculated.

MEP, motor evoked potential; Amp, amplitude; Lat, latency; TcMEP, transcranial motor evoked potential; SD, standard deviation; SEM, standard error of the mean.

Differences in symmetry of detection of muscles, transcranial motor evoked potential versus epidural motor stimulation

The bilateral stimulation of muscles is necessary for a neuromonitoring protocol. For TcMEP, the only completely reliable muscles for bilateral detection were the TA and ECR. For other muscles, the detection frequency varied from high to low, as indicated in Table 2 and Figure 3. ES increased the frequency of detection in several muscles, but was associated with more failures to record bilaterally than TcES. We calculated an index of asymmetric detection as the percent difference unilateral- bilateral detection × animal number per muscle. When summed across the tested muscles, the values are TcES = 846, T6 ES = 921.8, and T9 ES = 1028.4. Complete symmetry would be zero. Normalized to TcES, ES at T6 and T9, respectively, are associated with 1.1 and 1.2 greater asymmetry. This may reflect asymmetrical ES current flows attributed to slight deviations from midline, or perhaps more likely, induced inequality of the cerebrospinal fluid (CSF) space attributed to a greater pressure on one side of the dura than the other, with the effect to narrow the CSF space on one side and expand it on the opposite side.³²

FIG. 3.

Detection percentages of each muscle in 18 animals, considering at least one muscle present from left or right side. (A) Observe lower percentages of detection in vastus lateralis, rectus femoris, vastus medialis, and abductor hallucis using TcES, compared with ES at the T6 segment seen in (B) and T9 segment seen in (C). (D) Epidural stimulation elicits forelimb muscle (ECR) waveforms for 9/10 (T5/6) stimuli (89%) and 7/10 (T9/10) stimuli (72%). These stimuli elicited both short and longer latency MEPs with T9/10 stimulation generating a multi-spike long latency signal in ECR. (E) Distribution of latencies to several muscles in response to stimulation applied transcranially or with epidural electrodes. The modal values show the most frequently detected latencies among muscles. Note the bimodal distribution of the ECR in response to TcES. AH, abductor hallucis; BF, biceps femoris; ECR, extensor carpi radialis; ES, electrical stimulation; GAS, gastrocnemius; LDE, lateral digital extensor; RF, rectus femoris; VL, vastus lateralis; VM, vastus medialis; TA, tibialis anterior; TcES/TcES, transcranial electrical stimulation.

Motor evoked potential latency and amplitude differences

MEP latencies were more consistent than amplitudes (Supplementary Fig. 3) (see online supplementary material at http://www.liebertpub.com). TcES latencies were shorter to the forelimbs when compared to the hindlimbs (p < 0.0001) and longer than those generated with ES. ECR had the shortest TcES latency ( \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document}$$\overline {x}$$ \end{document} = 15.8 ± 5.2 ms SD). For hindlimb muscles, the shortest latency occurred in RF ( \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document}$$\overline {x}$$ \end{document} = 21.6 ± 6.7 ms SD) and the longest in AH ( \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document}$$\overline {x}$$ \end{document} = 32.2 ms ± 4.9 ms SD). T6 stimulation elicited hindlimb muscle responses with latency values of 15.6 ms for the RF and 23.6 ms for the AH. Stimulation over the T9 segment produced the shortest latency values to the hindlimbs (RF, 13.7 ms; LDE, 20.1 ms).

The mean latency difference between T6 and T9 ES stimulation for all muscles was 2.5 ± 0.5 ms SD, indicating the absence of iatrogenic cord injury attributed to the laminectomy. This short latency made it difficult to record signals in between the electrodes because of obscuration by the stimulation artifact. Possibly, the ES electrodes could be positioned further rostral and caudal to increase latency and facilitate cross-lesion conduction testing, but there is some risk of injury related to blindly advancing the paddle electrodes.

The latency values obtained in response to the three stimulation sites (T6, T9, and TcES) differed significantly (ANOVA; p < 0.001) for muscles with small SDs (AH, BF, Gas, and TA), but not for the large vastus medialis and lateralis muscles. Representative latencies between TcES and ES are shown in Figure 2D. TcES and ES latencies were fairly consistent among subjects with SDs generally less than 6 ms. The smallest latency SDs were in response to T9 ES.

MEP amplitude variability between animals was substantial with either TcMEPs or ES. ES at T6 and T9 generated less mean amplitude variability in TA and AH when compared to transcranial stimulation. The amplitude of TcES MEPs was greatest in the forelimbs extensors, ECR ( \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document}$$\overline {x}$$ \end{document} = 1655 ± 690 μV) being 3 × larger than TA amplitudes (TA: \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document}$$\overline {x}$$ \end{document} = 549 ± 281 μV SD; t-test p < 0.0001). Forelimb ECR had the smallest amplitude variability ( \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document}$$\overline {x}$$ \end{document} = 1655 μV ± 689 μV SD) in response to TcES. ES generated larger amplitudes in hindlimb TA and AH muscles, but not in other hindlimb muscles. Peak-to-peak amplitude variability was similar for all stimulation sites and cranial recording arrays. SD, as a fraction of amplitude across the detected muscles, was similarly high; 0.83 (TcES), 0.89 (T6), and 0.84 (T9).

Needle position had little effect on tibialis anterior motor evoked potentials

In 9 animals, we sought to determine whether MEPs are sensitive to the relative position of the electrodes within the most reproducible muscle (TA). We placed an extra pair of recording needle electrodes in a position lateral to the midline of the belly of the muscle (Fig. 1D). Signal was detected in all animals. Although the shape of the MEP was slightly different (Fig. 2E), there was no significant difference in latencies or amplitudes in between the two recording locations for TcES (t test; p > 0.05; Table 3). There was a small increment in amplitudes in response to ES for electrodes in the middle of the TA belly. In larger muscles, it may be necessary to seek the motor point and maintain the joint angles in a uniform position to reduce EMG detection and amplitude variability.

Table 3.

Recording Position Test

	Muscle midline position n = 18		Muscle lateral position n = 9
Stimulation modality	Amplitude uV TA2 uV (SD)	Latency ms TA2 ms (SD)	Amplitude uV TA1 (SD)	Latency ms TA1 (SD)
TcES	549 (281)	27.1 (4.6)	587 (290)	25.9 (3.9)
ES T6	1036 (742)	20.5 (3.9)	661 (468)	20.9 (4.1)
ES T9	1138 (670)	17.9 (3.4)	846 (690)	18.89 (4.1)

Two different recording electrode positions tested in the same tibialis anterior muscle.

TcES, transcranial electrical stimulation; ES, electrical stimulation; SD, standard deviation.

Forelimb extensor carpi radialis motor evoked potential in response thoracic epidural stimulation

ES at T6 and T9 elicited unexpected bilateral responses in forelimb ECR with two distinct latencies. ECR MEPs were detected for 89% of T6 ES stimuli and 56% from T9 (Table 2). The first MEP response had a mean latency from T6 of \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document}$$\overline {x}$$ \end{document} = 12.6 ± 3.1 ms (SD), and from T9, \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document}$$\overline {x}$$ \end{document} = 17.4 ± 4.6 ms (SD). A second long latency multiphasic waveform from T6 ES had a mean \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document}$$\overline {x}$$ \end{document} = 38.5 ms ± 6.3 ms (SD) in 8 animals (44%) and in 1 animal comprised the only signal recorded (37 ms; Fig. 3D). ES at T9 produced short latency MEPs (<25 ms) in forelimb ECRs in 9 animals. Late latency waveforms were observed in 6 animals ( \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document}$$\overline {x}$$ \end{document} = 38.8 ± 7.5 ms [SD]), 3 of which followed the early latency components. In another 3 animals, only the long latency component was detected.

Impact of progressive transection on somatosensory evoked potentials and motor evoked potentials

In the progressive T9 transection experiment, the first increment of dorsal incision slightly increased the latency, consistent with the intraoperative monitoring literature. When the full left dorsal hemisection was completed, the right cranial SSEP was lost (Fig. 5). The effect of this extent of lesion on the left TA MEP was an increase in latency and a 15 × reduction in amplitude. Thus, there must be some functioning motor pathways in the spared left spinal tissue after dorsal hemisection. Next, the right spinal cord was progressively incised until only a ventral bridge remained in continuity. This was associated with the persistence of small amplitude MEPs. This indicates that motor fibers in the ventral spinal cord convey MEPs. The absence of detected potentials after complete transection confirms that observed potentials are not artefactual.

FIG. 5.

Sequential spinal cord transection in a single animal; EPs were recorded before and after several transection steps. (A) SSEP recorded using the CP2–CP1 array in response to tibial nerve stimulation. (B) MEPs collected bilaterally from TA. Transection percentages shown in the schematic are estimates. After complete ipsilateral section of the dorsal column, the SSEP signal is lost, but the TA MEP is retained, although with reduced amplitude and increased latency. With further sectioning and only the right anterior funiculus spared, a small-amplitude MEP signal was collected in the ipsilateral TA. EPs, evoked potentials; LDC, ; MEPs, motor evoked potentials; SSEP, somatosensory evoked potential; TA, tibialis anterior;

Evoked potential procedure duration

After the electrophysiology assessments, each study animal underwent a thoracic contusion injury and was subsequently randomized to receive a cell transplant or control injection. Thus, the time required to perform the measures was important. Unlike established human neuromonitoring, these studies required troubleshooting and determination of the best stimulation and recording parameters. Insertion of the cranial screws for stimulation and recording and localization of the sciatic nerve was time-demanding as was recording from multiple muscles. If a laminectomy is performed, the insertion of epidural stimulating electrodes is very time efficient.

Table 4.

Is There a Spinal CST in the Yucatan Micropig?

Is there a spinal CST in the Yucatan micropig?
Evidence against	Evidence for
• Single-pulse stimuli unable to evoke upper limb MEPs.• Small cortical representation for limbs.• Older tract degeneration studies.	• Possible D waves recorded from T6.• MEPs evoked by intracortical stimulation.• CaM-kinase positive axons in lateral funiculus (unpublished).• TA MEP latency shorter than in humans.

CST, corticospinal tract; MEPS, motor evoked potentials; CaM, calcium and calmodulin; TA, tibialis anterior.

Discussion

We set out to develop intraoperative neuromonitoring techniques to aid our studies of SCI and transplantation in the Yucatan porcine model. Our findings refined our understanding of neurophysiological similarities and differences between humans and Yucatan minipigs. Further study is needed to clarify the cortical motor system and its projections. A general theme was that recordings to and from the upper extremity were more robust. Whether the generally smaller amplitudes for lower extremity recordings reflect differences in cortical representation, technical factors, or spinal tract anatomy requires further study.

Somatosensory evoked potential recordings

SSEPs were reliable measures. Median nerve and epidural stimulation evoked large amplitudes in all cortical arrays, allowing identification of peaks at P18 and N24. Tibial nerve stimulation produced smaller amplitude recordings with P29 and N35 peaks that were more susceptible to electrical noise. To determine whether sensory conduction is retained through an SCI, it may be better to stimulate from the sciatic nerve or the epidural space caudal to the injury site to improve the signal-to-noise ratio. Neutral (midline) referenced arrays had larger amplitudes than contralateral cortex referenced arrays, independent of the stimulation site. This was most evident for stimulation from the epidural thoracic segments. One explanation is a desynchronization effect because ES excites both sides of the spinal cord, and thus both cortexes are active at the same time during a CP1–CP2 recording.

The robust sensory stimulation observed with ES is attributed to the stimulation of more sensory fibers than can be excited from the tibial nerve. Ascending sensory fibers from the proximal leg and trunk below the stimulation level, as well as segmental nerve roots, may contribute to the SSEP. ES is thus more potent to generate cortical sensory signals, but has less specificity. In the sequential transection experiment, the detection of right cranial SSEPs was lost after left dorsal column transection. This indicates that the SSEP signal is carried in a dorsally located pathway.

The latency values found in the Yucatan micropig in response to peripheral nerve stimulation are similar to values reported in other large animal models.^12,33 Owen and colleagues reported a 35-ms tibial-evoked cortical SSEP in the hog. Although those are larger pigs, this value is within the range of our measurements.

Motor neuromonitoring

The most reliable muscles for TcES monitoring were unknown to us, so we tested the responsiveness of nine pairs of limb muscles. TA was the most consistently detected muscle in response to either transcranial or epidural stimulation. Recording from a “reliable” muscle is useful to address technical factors to optimize the recordings. Tibialis anterior and abductor hallucis MEPs are used in human clinical intraoperative monitoring protocols because of their reliability.³⁴ AH was not as consistently detected in this study. In response to TcES, amplitudes recorded in the forelimb (ECR) MEPs were larger as compared to hindlimb muscles, providing a reliable control measure.

In awake humans, the latency of TA recordings in response to transcranial magnetic stimulation (TMS) is approximately 30 ms in anesthetized patients.^35,36 TA latency values in anesthetized humans using TcES delivered through screw electrodes are slightly longer (mean, 35 ms; SD, 3.8 ms).³⁷ In the pigs, mean TA latency after TcES was 27 ms, compatible with the shorter overall length of the pigs (approximately 105 cm).

Given the lack of knowledge regarding motor and sensory pathways in the pig, we considered the possibility that ES current spread to a thoracic corticospinal tract (CST) could produce an antidromic signal going directly to the motor cortex without intervening synapses,³⁸ confounding the interpretation of epidural SSEPs (arising in the dorsal columns). Antidromic stimulation of motor fibers is used to measure motor axon conduction velocity.³⁹ We regard this possibility as an unlikely source of SSEP contamination because the latency of the cortical sensory response from T6 ES was 16–17 ms. SSEPs have at least two intervening synapses: in the gracilis nucleus and the thalamus. TA MEP latencies were used to estimate the time required for antidromic motor tract transmission. We subtracted the latency of T6 ES stimulation (20.5 ms) from the latency of TcES (27.1 ms) stimulation. This 6.6-ms difference does not match the SSEP latency.

Epidural versus cranial motor stimulation

ES elicited MEPs in more hindlimb muscles than cortical stimulation, and with larger amplitudes. MEPs from RF, LDE, and VL, infrequently detected from TcES (<20% per muscle), were readily detected using thoracic ES (80% at T6 and 90% at T9). ES appears to recruit more motor axons and tracts than cranial stimulation. The observation may also reflect that large muscles close to the body axis are mainly innervated by brainstem reticulo- or rubrospinal tracts in quadrupeds.^40

–43

Epidural stimulation leads to motor evoked potentials in forelimb muscles

ES at T6 and T9 elicited an MEP in ECR, the only monitored forelimb muscle indicating the presence of axons that project from or through these thoracic levels to cervical motoneurons. These axons may be long ascending propriospinal axons extending from the lumbosacral enlargement involved in the forelimb-hindlimb coordination of quadrupedal locomotion, and are of interest in recovery from SCI.^44,45 The ECR MEPs evoked by T6 and T9 ES had distinct short and long latency waveforms (Table 1). The shorter latency may be attributed to more direct activation of axons terminating on ECR motoneurons. The longer latency potentials had a multi-spike appearance, in keeping with a multi-synaptic origin. Also notable is that T6 and T9 ES evoked MEPs simultaneously from both upper and lower extremity muscles, as shown in Figure 4 (bilateral ECR and TA). This multi-segmental motor stimulation³⁰ may allow ES to be valuable to study propriospinal plasticity and repair post-injury. It would be of interest in subsequent studies to determine whether lower extremity peripheral nerve stimulation could trigger ECR MEPs, as reported by Calancie, in persons with chronic complete SCI.⁴⁶

FIG. 4.

Forelimb and hindlimb MEPs after thoracic epidural stimulation. Five MEP samples collected simultaneously from left and right ECR and TA in response to ES, first, at T6 and then at the T9/10 segment. As expected, the caudal lead evokes shorter latencies in the leg muscles compared to the T6 segment. However, in the forelimb muscle there is a 7-ms difference in latencies evoked from the epidural T9 position when compared to T6 suggesting differing pathway activation (see Discussion). The difference in amplitudes between sides suggests a deviation of the epidural lead contacts towards the left. ECR, extensor carpi radialis; ES, electrical stimulation; MEPs, motor evoked potentials; TA, tibialis anterior.

Is there a corticospinal tract in Yucatan minipigs?

Evidence for a spinal CST in the pig is lacking. Older studies suggested that the CST did not descend below the upper cervical level,¹⁹ having mainly corticobulbar terminations. In the cat, cortical input to the reticulospinal system has been shown to mediate transmission to motor neurons of the spinal cord.⁴⁷ To determine whether a spinal CST is present, modern neuroanatomic tracing, neurophysiology, and imaging techniques, such as diffusion tensor imaging, will be useful.⁴⁸ Given the association between a large pyramidal tract and the evolution of hand function in primate species, a spinal CST may not be an important source of motor control in hooved quadruped animals.

Whether or not there is a spinal CST, we obtained motor transmission from electrical excitation near the cortical surface; thus, motor neuromonitoring can be used in these animals.

The location and topography of the motor cortex also requires further study. The great majority of cortex is devoted to the snout.²⁰ Apparently, there is no separate motor cortical gyrus; the cruciate gyrus includes both motor and somatosensory cortex without sulcal delineation. The hindlimb representation is described to reside within the interhemispheric wall of the cruciate gyrus deep to overlying cortex (see diagram, Supplementary Figs. 1 and 2) (see online supplementary material at http://www.liebertpub.com). The forelimb representation was described to occupy a more cranial position in the lateral wall and base of the cruciate sulcus.^{18
–20,49,50} The few reports that have used direct cortical stimulation to localize the primary motor area reported that trains of at least three pulses were required to elicit motor activity in forelimbs and hindlimbs.^18,19 Motor conduction from the cortex through more than one synapse to activate alpha motor neurons would explain the need of multiple pulses and the high sensitivity to anesthetics.

A small and deeply located limb representation may also account for the high stimulation intensities required. We conducted a localization pilot study in 3 animals using direct cortical stimulation while recording EMG. The purpose was to localize the forelimb and hindlimb regions for injection of anterograde tracers (see Supplementary Fig. 1) (see online supplementary material at http://www.liebertpub.com). Upper extremity MEPS were triggered by stimuli in the coronal gyrus on the medial cortical surface. Lower extremity MEPs were evoked by stimuli deep to the surface within the cruciate gyrus of the interhemispheric fissure.

Another technique to identify CST conduction is to record D waves from dorsally placed epidural electrodes within the spinal canal in response to single-pulse cortical stimulation. D waves are “direct” axonal potentials without intervening synapses. Typically, this conduction occurs in the CST, but, theoretically, D waves might arise from other directly stimulated tracts. I waves are generated attributed to inhibitory signaling within the motor cortex and confirm that the observed excitation is cortical. We attempted to record D waves in 3 animals using an epidural electrode placed at T6. In Supplementary Figure 4 (see online supplementary material at http://www.liebertpub.com), we show spinal cord waveforms detected with a short latency after single or double TcES pulses. A D wave is present. The after-following waves observed may represent D wave trifurcation, as reported by Macdonald, or, less possibly, I waves.⁵¹ This finding indicates a direct connection from the stimulated region to the spinal level. Direct and indirect waves will be assessed in subsequent studies using cortical surface stimulation to ensure a more limited depth of current. If a spinal CST is present, D waves could be useful markers of injury and recovery. Although we do not yet having tracing data, we have found bundles of fibers in the lateral funiculus at C1 that stained with Cam-Kinase II alpha antibody (unpublished data). This enzyme has been identified within CST fibers of rodent species.⁵²

High-voltage TcES generated a bimodal distribution of ECR latencies at 10 and 20 ms in some animals. High-voltage stimulation may have reached subcortical levels, and the different MEP latencies may reflect transmission in different pathways. The contribution of multiple tracts to MEPs in response to TMS has been characterized in the rat, where the responses had a similar latency dispersion to the bimodal distribution we found.³⁴

The progressive transection experiment was suggestive that both dorsolateral and ventral motor pathways could be stimulated with TcES. Dorsolateral pathways may include the CST and rubrospinal, and ventral pathways the reticulospinal and CST.

Sources of variability in the study

MEP variability is widely reported in the literature, including similar porcine models.^53
–55 Some inconsistency is attributed to technical factors, but the age, size, anatomy, and neurophysiology of the animals may also contribute.⁵³ Intraoperative EPs are markedly influenced by level of anesthesia, rate of drug metabolism, electrode impedance, monitoring noise, body temperature, blood pressure, spinal cord blood flow, and motor neuron excitability.^51,52 Maintenance of core temperatures is challenging under general anesthesia; we often had to supplement our heating methods to correct low temperature levels.

Anesthetics may contribute to intersubject variability; ketamine increases MEP amplitude values and intersubject variability in humans.⁵⁶ Another possible source of variation may be incomplete development and myelination of central nervous system pathways.³⁰ The age of the pigs used in the studies is largely based on cost and husbandry issues that favor younger smaller animals. We are conducting histological studies to assess whether tracts show mature myelination in these juvenile animals.

Despite reducing impedance with skull seated screw electrodes, it was still necessary to apply high voltages and multiple pulses to generate TcES-MEPs. As shown in Supplementary Figure 2 (see online supplementary material at http://www.liebertpub.com), there are substantial interanimal differences in cortical morphology. Thus, the relative distribution of the motor and sensory cortex to the location of screw electrodes may vary. The larger MEPs observed in the forelimb might reflect the relatively anterior screw position and a more superficial location of the forelimb cortex. In future animals, we will explore optimizing the electrode position based on amplitudes rather than external landmarks. Another anatomical issue is that the Yucatan micropigs have an extensive frontal sinus that enlarges with age.^57,58 If a screw is within or close to a sinus, the impedance may be elevated and the infection risk increased.

Regarding ES, we could not visualize the epidural electrode contacts after they were placed under the lamina, and they may not always be optimally positioned to stimulate through the dura. Our electrodes were relatively stiff, and thus the contacts may not be fully seated to optimize conductance. Further, the thickness of the CSF layer in the Yucatan subdural space varies between spinal levels and animals (unpublished magnetic resonance imaging and ultrasound data) affecting current distribution. These factors might also account for the detection of only a unilateral MEP detection in some muscles.

Neuromonitoring protocol

In summary, an intraoperative electrophysiological protocol was created for the Yucatan model, the most reliable muscles to be monitored were established, and reference ranges determined. A basic neuromonitoring protocol would consist of median and tibial nerve SSEPs, and MEPs recorded from ECR in the upper and TA in the lower extremities. Taking extra time to localize the sciatic nerves could provide larger SSEP amplitudes from the legs, increasing sensitivity. Motor and sensory epidural stimulation spanning the planned contusion site will yield even larger amplitudes to detect preserved transmission from above and below the injury. Further, this methodology will be useful to isolate the transmission impairment to the injury site.

Important questions for future studies are to define the pathways that mediate quadruped locomotion and optimize stimulation techniques. The utility of the model to test therapeutics intended for human use is linked to whether similar systems contribute to motor function between Yucatan micropigs and humans. The model may offer insights into the role of propriospinal circuits in regulating motor function.

Footnotes

Acknowledgments

This work would not have been possible without the generous support of The Miami Project to Cure Paralysis and The Buoniconti Fund, and by grants from The State of Florida Department of Health.

Author Disclosure Statement

No competing financial interests exist.

References

Curt

, and Ellaway

P.H.

(2012). Clinical neurophysiology in the prognosis and monitoring of traumatic spinal cord injury. Handb. Clin. Neurol., 109, 63–75.

, Houlden

D.A.

, and Rowed

D.W.

(1990). Somatosensory evoked potentials and neurological grades as predictors of outcome in acute spinal cord injury. J. Neurosurg., 72, 600–609.

Biering-Sorensen

, Alai

, Anderson

, Charlifue

, Chen

, DeVivo

, Flanders

A.E.

, Jones

, Kleitman

, Lans

, Noonan

V.K.

, Odenkirchen

, Steeves

, Tansey

, Widerstrom-Noga

, and Jakeman

L.B.

(2015). Common data elements for spinal cord injury clinical research: a National Institute for Neurological Disorders and Stroke project. Spinal Cord, 53, 265–277.

John

E.R.

, Chabot

R.J.

, Prichep

L.S.

, Ransohoff

, Epstein

, and Berenstein

(1989). Real-time intraoperative monitoring during neurosurgical and neuroradiological procedures. J. Clin. Neurophysiol., 6, 125–158.

Deletis

, and Sala

(2008). Intraoperative neurophysiological monitoring of the spinal cord during spinal cord and spine surgery: a review focus on the corticospinal tracts. Clin. Neurophysiol., 119, 248–264.

Deletis

(2007). Basic methodological principles of multimodal intraoperative monitoring during spine surgeries. Eur. Spine J., 16, Suppl. 2, S147–S152.

Niimi

, Sala

, Deletis

, Setton

, de Camargo

A.B.

, and Berenstein

(2004). Neurophysiologic monitoring and pharmacologic provocative testing for embolization of spinal cord arteriovenous malformations. AJNR Am. J. Neuroradiol., 25, 1131–1138.

Holdefer

R.N.

, MacDonald

D.B.

, and Skinner

S.A.

(2015). Somatosensory and motor evoked potentials as biomarkers for post-operative neurological status. Clin. Neurophysiol., 126, 857–865.

Lee

J.H.

, Jones

C.F.

, Okon

E.B.

, Anderson

, Tigchelaar

, Kooner

, Godbey

, Chua

, Gray

, Hildebrandt

, Cripton

, Tetzlaff

, and Kwon

B.K.

(2013). A novel porcine model of traumatic thoracic spinal cord injury. J. Neurotrauma, 30, 142–159.

10.

Watanabe

, Andersen

, Simonsen

C.Z.

, Evans

S.M.

, Gjedde

, Cumming

, and DaNe

X.S.G.

(2001). MR-based statistical atlas of the Gottingen minipig brain. Neuroimage, 14, 1089–1096.

11.

Yun

S.P.

, Kim

D.H.

, Ryu

J.M.

, Park

J.H.

, Park

S.S.

, Jeon

J.H.

, Seo

B.N.

, Kim

H.J.

, Park

J.G.

, Cho

K.O.

, and Han

H.J.

(2011). Magnetic resonance imaging evaluation of Yukatan minipig brains for neurotherapy applications. Lab. Anim. Res., 27, 309–316.

12.

Owen

J.H.

, Jenny

A.B.

, Naito

, Weber

, Bridwell

K.H.

, and McGhee

(1989). Effects of spinal cord lesioning on somatosensory and neurogenic-motor evoked potentials. Spine, 14, 673–682.

13.

Bjarkam

C.R.

, Glud

A.N.

, Margolin

, Reinhart

, Franklin

, Deding

, Ettrup

K.S.

, Fitting

L.M.

, Nielsen

M.S.

, Sørensen

J.C.

, and Cunningham

M.G.

(2010). Safety and function of a new clinical intracerebral microinjection instrument for stem cells and therapeutics examined in the Göttingen minipig. Stereotact. Funct. Neurosurg., 88, 56–63.

14.

Guest

, Benavides

, Padgett

, Mendez

, and Tovar

(2011). Technical aspects of spinal cord injections for cell transplantation. Clinical and translational considerations. Brain Res. Bull., 84, 267–279.

15.

Federici

, Hurtig

C.V.

, Burks

K.L.

, Riley

J.P.

, Krishna

, Miller

B.A.

, Sribnick

E.A.

, Miller

J.H.

, Grin

, Lamanna

J.J.

, and Boulis

N.M.

(2012). Surgical technique for spinal cord delivery of therapies: demonstration of procedure in gottingen minipigs. J. Vis. Exp. (70), e4371.

16.

Procter

L.D.

, Davenport

D.L.

, Bernard

A.C.

, and Zwischenberger

J.B.

(2010). General surgical operative duration is associated with increased risk-adjusted infectious complication rates and length of hospital stay. J. Am. Coll. Surg., 210, 60–65.e1–2.

17.

Kim

B.D.

, Hsu

W.K.

, De Oliveira

G.S.

Jr ., Saha

, and Kim

J.Y.

(2014). Operative duration as an independent risk factor for postoperative complications in single-level lumbar fusion: an analysis of 4588 surgical cases. Spine, 39, 510–520.

18.

Thompson

W.D.

(1966). Study of the motor cortex of the domestic pig. Am. J. Vet. Res., 27, 1369–1373.

19.

Palmieri

, Farina

, Panu

, Asole

, Sanna

, De Riu

P.L.

, and Gabbi

(1986). Course and termination of the pyramidal tract in the pig. Arch. Anat. Microsc. Morphol. Exp., 75, 167–176.

20.

Craner

S.L.

, and Ray

R.H.

(1991). Somatosensory cortex of the neonatal pig: I. Topographic organization of the primary somatosensory cortex (SI). J. Comp. Neurol., 306, 24–38.

21.

Toleikis

J.R.

; American Society of Neurophysiological Monitoring. (2005). Intraoperative monitoring using somatosensory evoked potentials. A position statement by the American Society of Neurophysiological Monitoring. J. Clin. Monit. Comput., 19, 241–258.

22.

Szelenyi

, Kothbauer

K.F.

, and Deletis

(2007). Transcranial electric stimulation for intraoperative motor evoked potential monitoring: Stimulation parameters and electrode montages. Clin. Neurophysiol., 118, 1586–1595.

23.

Lyon

, Gibson

, Burch

, and Lieberman

(2010). Increases in voltage may produce false-negatives when using transcranial motor evoked potentials to detect an isolated nerve root injury. J. Clin. Monit. Comput., 24, 441–448.

24.

Sloan

T.B.

, and Heyer

E.J.

(2002). Anesthesia for intraoperative neurophysiologic monitoring of the spinal cord. J. Clin. Neurophysiol., 19, 430–443.

25.

Ertekin

, Sarica

, and Uckardesler

(1984). Somatosensory cerebral potentials evoked by stimulation of the lumbo-sacral spinal cord in normal subjects and in patients with conus medullaris and cauda equina lesions. Electroencephalogr. Clin. Neurophysiol., 59, 57–66.

26.

Beric

, Dimitrijevic

M.R.

, Sharkey

P.C.

, and Sherwood

A.M.

(1986). Cortical potentials evoked by epidural stimulation of the cervical and thoracic spinal cord in man. Electroencephalogr. Clin. Neurophysiol., 65, 102–110.

27.

Zentner

(1989). Scalp recorded somatosensory evoked potentials in response to cauda equina stimulation in neurosurgical operations on the spinal cord. Br. J. Neurosurg., 3, 39–43.

28.

Paradiso

, De Vito

, Rossi

, Setacci

, Battistini

, Cioni

, Passero

, Giannini

, and Rossini

P.M.

(1995). Cervical and scalp recorded short latency somatosensory evoked potentials in response to epidural spinal cord stimulation in patients with peripheral vascular disease. Electroencephalogr. Clin. Neurophysiol., 96, 105–113.

29.

Wolter

, Kieselbach

, Sircar

, and Gierthmuehlen

(2013). Spinal cord stimulation inhibits cortical somatosensory evoked potentials significantly stronger than transcutaneous electrical nerve stimulation. Pain Physician, 16, 405–414.

30.

Taylor

B.A.

, Fennelly

M.E.

, Taylor

, and Farrell

(1993). Temporal summation–the key to motor evoked potential spinal cord monitoring in humans. J. Neurol. Neurosurg. Psychiatry, 56, 104–106.

31.

Calancie

, Klose

K.J.

, Baier

, and Green

B.A.

(1991). Isoflurane-induced attenuation of motor evoked potentials caused by electrical motor cortex stimulation during surgery. J. Neurosurg., 74, 897–904.

32.

Struijk

J.J.

, Holsheimer

, van Veen

B.K.

, and Boom

H.B.

(1991). Epidural spinal cord stimulation: calculation of field potentials with special reference to dorsal column nerve fibers. IEEE Trans. Biomed. Eng., 38, 104–110.

33.

Wilson

R.D.

, and Beerwinkle

K.R.

(1986). Somatosensory-evoked potential induced by stimulation of the caudal tibial nerve in awake and barbiturate-anesthetized sheep. Am. J. Vet. Res., 47, 46–49.

34.

Y.L.

, Dan

Y.F.

, Tan

Y.E.

, Nurjannah

, Tan

S.B.

, Tan

C.T.

, and Raman

(2006). Intraoperative motor-evoked potential monitoring in scoliosis surgery: comparison of desflurane/nitrous oxide with propofol total intravenous anesthetic regimens. J. Neurosurg. Anesthesiol., 18, 211–214.

35.

Kawaguchi

, Sakamoto

, Inoue

, Kakimoto

, Furuya

, Morimoto

, and Sakaki

(2000). Low dose propofol as a supplement to ketamine-based anesthesia during intraoperative monitoring of motor-evoked potentials. Spine, 25, 974–979.

36.

Furby

, Bourriez

J.L.

, Jacquesson

J.M.

, Mounier-Vehier

, and Guieu

J.D.

(1992). Motor evoked potentials to magnetic stimulation: technical considerations and normative data from 50 subjects. J. Neurol., 239, 152–156.

37.

Watanabe

, Watanabe

, Takahashi

, Saito

, Hirato

, and Sasaki

(2004). Transcranial electrical stimulation through screw electrodes for intraoperative monitoring of motor evoked potentials. Technical note. J. Neurosurg., 100, 155–160.

38.

Partanen

, Merikanto

, Kokki

, Kilpelainen

, and Koistinen

(2000). Antidromic corticospinal tract potential of the brain. Clin. Neurophysiol., 111, 489–495.

39.

Mediratta

N.K.

, and Nicoll

J.A.

(1983). Conduction velocities of corticospinal axons in the rat studied by recording cortical antidromic responses. J. Physiol., 336, 545–561.

40.

Calancie

, Molano

M.R.

, and Broton

J.G.

(2004). Abductor hallucis for monitoring lower-limb recovery after spinal cord injury in man. Spinal Cord, 42, 573–580.

41.

Drew

, Jiang

, and Widajewicz

(2002). Contributions of the motor cortex to the control of the hindlimbs during locomotion in the cat. Brain research. Brain Res. Rev., 40, 178–191.

42.

Calancie

, Alexeeva

, Broton

J.G.

, Suys

, Hall

, and Klose

K.J.

(1999). Distribution and latency of muscle responses to transcranial magnetic stimulation of motor cortex after spinal cord injury in humans. J. Neurotrauma, 16, 49–67.

43.

Isa

, Kinoshita, M. and Nishimura

(2013). Role of direct vs. indirect pathways from the motor cortex to spinal motoneurons in the control of hand dexterity. Front. Neurol., 4, 191.

44.

Miller

, Reitsma

D.J.

, and van der Meche

F.G.

(1973). Functional organization of long ascending propriospinal pathways linking lumbo-sacral and cervical segments in the cat. Brain Res. 62, 169–188.

45.

Flynn

J.R.

, Graham

B.A.

, Galea

M.P.

, and Callister

R.J.

(2011). The role of propriospinal interneurons in recovery from spinal cord injury. Neuropharmacology, 60, 809–822.

46.

Calancie

(1991). Interlimb reflexes following cervical spinal cord injury in man. Exp. Brain Res., 85, 458–469.

47.

Alstermark

, Pinter

, and Sasaki

(1983). Brainstem relay of disynaptic pyramidal EPSPs to neck motoneurons in the cat. Brain Res. 259, 147–150.

48.

Dyrby

T.B.

, Sogaard

L.V.

, Parker

G.J.

, Alexander

D.C.

, Lind

N.M.

, Baare

W.F.

, Hay-Schmidt

, Eriksen

, Pakkenberg

, Paulson

O.B.

, and Jelsing

(2007). Validation of in vitro probabilistic tractography. Neuroimage, 37, 1267–1277.

49.

Saikali

, Meurice

, Sauleau

, Eliat

P.A.

, Bellaud

, Randuineau

, Verin

, and Malbert

C.H.

(2010). A three-dimensional digital segmented and deformable brain atlas of the domestic pig. J. Neurosci. Methods, 192, 102–109.

50.

Sauleau

, Lapouble

, Val-Laillet

, and Malbert

C.H.

(2009). The pig model in brain imaging and neurosurgery. Animal, 3, 1138–1151.

51.

Macdonald

D.B.

(2006). Intraoperative motor evoked potential monitoring: overview and update. J. Clin. Monit. Comput., 20, 347–377.

52.

Terashima

(1995). Anatomy, development and lesion-induced plasticity of rodent corticospinal tract. Neurosci. Res., 22, 139–161.

53.

Burke

, Hicks

, Stephen

, Woodforth

, and Crawford

(1995). Trial-to-trial variability of corticospinal volleys in human subjects. Electroencephalogr. Clin. Neurophysiol., 97, 231–237.

54.

Burke

and Hicks

(1995). Intraoperative monitoring of corticospinal function. Electroencephalogr. Clin. Neurophysiol., 94, 89–91.

55.

Mok

J.M.

, Lyon

, Lieberman

J.A.

, Cloyd

J.M.

, and Burch

(2008). Monitoring of nerve root injury using transcranial motor-evoked potentials in a pig model. Spine, 33, E465–E473.

56.

Inoue

, Kawaguchi

, Kakimoto

, Sakamoto

, Kitaguchi

, Furuya

, Morimoto

, and Sakaki

(2002). Amplitudes and intrapatient variability of myogenic motor evoked potentials to transcranial electrical stimulation during ketamine/N₂O- and propofol/N₂O-based anesthesia. J. Neurosurg. Anesthesiol., 14, 213–217.

57.

Hillman

D.J.

(1971). Macroscopic anatomy of the nasal cavity andparanasal sinuses of the domestic pig (Sus scrofa domestica). Retrospective Theses and Dissertations. Iowa State University Digital repositary. Paper 1160.

58.

Honig

J.F.

, Merten

H.A.

, Schutte

, Grohmann

U.A..

and Cassisis

(2002). Experimental study of the frontal sinus development on Goettingen miniature pigs. J. Craniofac. Surg., 13, 418–426.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.07 MB

0.20 MB

0.11 MB