Evaluating the Efficacy of Chronic Galantamine on Sustained Attention and Cholinergic Neurotransmission in A Pre-Clinical Model of Traumatic Brain Injury

Abstract

Cholinergic disruptions underlie attentional deficits following traumatic brain injury (TBI). Yet, drugs specifically targeting acetylcholinesterase (AChE) inhibition have yielded mixed outcomes. Therefore, we hypothesized that galantamine (GAL), a dual-action competitive AChE inhibitor and α7 nicotinic acetylcholine receptor (nAChR) positive allosteric modulator, provided chronically after injury, will attenuate TBI-induced deficits of sustained attention and enhance ACh efflux in the medial prefrontal cortex (mPFC), as assessed by in vivo microdialysis. In Experiment 1, adult male rats (n = 10–15/group) trained in the 3-choice serial reaction time (3-CSRT) test were randomly assigned to controlled cortical impact (CCI) or sham surgery and administered GAL (0.5, 2.0, or 5.0 mg/kg; i.p.) or saline vehicle (VEH; 1 mL/kg; i.p) beginning 24-h post-surgery and once daily thereafter for 27 days. Measures of sustained attention and distractibility were assessed on post-operative days 21–25 in the 3-CSRT, following which cortical lesion volume and basal forebrain cholinergic cells were quantified on day 27. In Experiment 2, adult male rats (n = 3–4/group) received a CCI and 24 h later administered (i.p.) one of the three doses of GAL or VEH for 21 days to quantify the dose-dependent effect of GAL on in vivo ACh efflux in the mPFC. Two weeks after the CCI, a guide cannula was implanted in the right mPFC. On post-surgery day 21, baseline and post-injection dialysate samples were collected in a temporally matched manner with the cohort undergoing behavior. ACh levels were analyzed using reverse phase high-performance liquid chromatography (HPLC) coupled to an electrochemical detector. Cortical lesion volume was quantified on day 22. The data were subjected to ANOVA, with repeated measures where appropriate, followed by Newman–Keuls post hoc analyses. All TBI groups displayed impaired sustained attention versus the pooled SHAM controls (p’s < 0.05). Moreover, the highest dose of GAL (5.0 mg/kg) exacerbated attentional deficits relative to VEH and the two lower doses of GAL (p’s < 0.05). TBI significantly reduced cholinergic cells in the right basal forebrain, regardless of treatment condition, versus SHAM (p < 0.05). In vivo microdialysis revealed no differences in basal ACh in the mPFC; however, GAL (5.0 mg/kg) significantly increased ACh efflux 30 min following injection compared to the VEH and the other GAL (0.5 and 2.0 mg/kg) treated groups (p’s < 0.05). In both experiments, there were no differences in cortical lesion volume across treatment groups (p’s > 0.05). In summary, albeit the higher dose of GAL increased ACh release, it did not improve measures of sustained attention or histopathological markers, thereby partially supporting the hypothesis and providing the impetus for further investigations into alternative cholinergic pharmacotherapies such as nAChR positive allosteric modulators.

Introduction

In the United States, approximately 5.3 million people live with long-term traumatic brain injury (TBI)-related disabilities, including functional, cognitive, and affective impairments.¹ Attentional processes, such as the ability to attend to, direct, and select relevant stimuli, are central to many basic and higher-order cognitive functions, with TBI-induced impairments affecting rehabilitative efforts, return to school and work, and overall quality of life.^2,3 Currently, there are no FDA-approved first-line pharmacotherapies to remediate attentional deficits following TBI.

Attentional functioning relies on an intact cholinergic network from the basal forebrain (BF) to the medial prefrontal cortex (mPFC).^4–7 Several lines of evidence suggest that these regions are particularly vulnerable to TBI, as evidenced by transient pathogenic changes in choline acetyltransferase (ChAT) and acetylcholinesterase (AChE), key enzymes involved in acetylcholine (ACh) synthesis and catabolism, respectively.^8–12 Moreover, chronic downregulation of ChAT is suggested to impair ACh bioavailability; thus, pharmacological agents that enhance cholinergic tone are hypothesized to improve neurobehavioral recovery.^8,9 Yet, selective AChE inhibitors such as donepezil, which increase the intrasynaptic availability of ACh, have mixed results and at high dosages may even be deleterious to TBI recovery.^13–15 Alternatively, enhancing cholinergic neurotransmission in the brain with drugs such as galantamine (GAL), a dual-action competitive AChE inhibitor and nicotinic receptor positive allosteric modulator, may extend greater therapeutic efficacy to mitigate complex attention deficits after TBI.

Several pre-clinical studies have demonstrated that GAL administration, using acute and chronic dosing paradigms, improves cognitive and histopathological deficits after TBI. Acute GAL injections (i.e., 30 min post-injury, twice daily for 3 days) improved learning and memory, reduced blood–brain barrier permeability, and attenuated GABAergic cell loss and newborn neuron loss in the ipsilateral hippocampus.¹⁶ Daily administration of GAL also enhanced spatial learning and memory in a water maze test whether tested alone or when combined with a subtherapeutic dose of environmental enrichment.^17,18 In addition, Njoku and colleagues showed that daily GAL treatment for 4 weeks post-injury enhanced cognitive flexibility in a PFC-dependent attentional set-shifting task and led to a dose-dependent reduction in cortical cavitation.¹⁹ These findings highlight the potential of GAL to serve as an adjunct therapy following TBI by enhancing cognitive performance and providing neuroprotection through histopathological mechanisms. Therefore, additional studies evaluating the effects of GAL using complex, complementary domains of cognitive functioning such as sustained attention, along with the novel assessment of the pharmacodynamics of GAL in the injured brain, will elucidate the therapeutic potential and neurochemical changes associated with pro-cholinergic pharmacological interventions after TBI.

Hence, the aim of this study was to utilize a two-phase approach to evaluate the effectiveness of three GAL doses on behavioral assessments, histological analysis, and in vivo microdialysis sampling methods. The first experiment examined the dose-dependent effects of chronic GAL treatment on restoring sustained attention performance in a 3-choice serial reaction time (3-CSRT) test, in which we previously reported injury-related impairments, as well as histopathological cholinergic integrity after TBI.²⁰ Utilizing the same experimental timeline, the second experiment investigated the pharmacodynamics of the three doses of GAL on drug-evoked in vivo ACh efflux in the mPFC in awake and freely-moving injured versus noninjured rats. It was hypothesized that chronic GAL administration post-injury would enhance attention measures and restore both qualitative (i.e., histopathology) and quantitative (i.e., in vivo microanalysis) measures of cholinergic tone in a dose-dependent manner.

Methods

Animals and pre-surgical procedures

The study is composed of two distinct experiments as follows (Fig. 1): Experiment 1 examines the therapeutic efficacy of GAL relative to saline vehicle (VEH) on post-injury cognitive and histopathological outcomes and Experiment 2 examines evoked ACh release through microdialysis in the mPFC following the administration of GAL or VEH. Eighty-four age-matched adult male Sprague-Dawley rats (Envigo RMS, Inc., Indianapolis, IN) were utilized (n = 68, Experiment 1 and n = 16, Experiment 2). The rats were pair-housed in standard laboratory ventilated polycarbonate cages within a temperature and light-controlled vivarium (12-h light/12-h dark; lights on at 7:00 a.m.) with ad libitum access to food and water. In both experiments, the rats underwent mild food restriction to maintain 80% of their body weight while having free access to water. All procedures were approved by the Institutional Animal Care and Use Committee at the University of Pittsburgh with every effort made to minimize the number of rats used and to alleviate pain and discomfort.

FIG. 1.

Schematic representation of experimental paradigms. In Experiment 1 (top panel), adult male rats were trained in the 3-choice serial reaction time (3-CSRT) test, and upon meeting criterion performance was randomly assigned to either a controlled cortical impact (CCI) or sham surgery. Galantamine (GAL; 0.5, 2.0, or 5.0 mg/kg) or vehicle (VEH; 1 mL/kg) was administered starting 24-h after CCI or Sham surgery and continued once daily for 27 days. On behavioral days, GAL or VEH was administered 30 min before behavior onset. Three weeks after surgery, rats were tested on the 3-CSRT for 5 days. In Experiment 2 (bottom panel), age-matched rats were given a CCI surgery, administered the same doses of GAL or VEH as in Experiment 1, but from days 1–22 and 14 days later a guide cannula was surgically implanted in the medial prefrontal cortex (mPFC). One week after the guide cannula insertion (i.e., 3 weeks post-injury), a microdialysis probe was inserted into the mPFC and baseline and post-injection dialysate was collected.

Surgery

Experiment 1: Controlled cortical impact (CCI) or sham surgery was performed as previously described.^{14,15,17–22} Briefly, surgical anesthesia was induced and maintained with 4% and 2% isoflurane, respectively, in a 2:1 N₂O:O₂ mixture. The rats were subsequently secured in a stereotaxic frame, and a craniectomy was made in the right hemisphere encompassing bregma, lambda, and the sagittal and coronal sutures. A moderate CCI was produced by impacting the exposed parietal cortex 2.8 mm with a 6 mm diameter impactor tip at a velocity of 4 m/s over the parietal cortex. Sham-operated rats underwent all surgical procedures sans the impact.

Experiment 2 consisted of a two-stage surgical procedure that prepared the rats for CCI injury and probe implantation for in vivo microdialysis. In the first stage, all rats received the CCI procedure described above, and in the second phase, which took place 2 weeks after CCI, a guide cannula was implanted into the right mPFC under the same anesthetic regimen described for Experiment 1. Specifically, while secured in a stereotaxic frame, four stainless steel bone anchor screws (Harvard Apparatus) were secured to the skull, with two on the left parietal skull and two on the frontal skull plate. A burr hole was made with a dental drill using an 8-degree lateral approach at coordinates AP +2.6 mm and ML +1.1 mm from bregma. Subsequently, the microdialysis guide cannula (SciPro, MAB 6.9.IC) was inserted through the burr hole using a stereotaxic arm and lowered to a depth of DV −3.3 mm below the dura and securely affixed to the skull using dental acrylic (Koldmount^TM). Stereotaxic coordinates were derived from the rat atlas of Paxinos and Watson (1986).²³ In addition, a tether screw was embedded in the dental acrylic posterior to lambda for tether connection. Core body temperature was monitored with a rectal thermistor and maintained at 37 ± 0.5°C using a heating blanket throughout the surgical procedures.

Acute neurological assessments

Following cessation of anesthesia, limb reflex ability was determined by gently squeezing the left and right hind paw every 5 s and recording the time to provoke a withdrawal response. Righting reflex was assessed by measuring the time required to turn from the supine to prone position over three consecutive trials.

Drug administration

In both experiments, GAL (Sigma-Aldrich, St. Louis, MO) was made fresh daily by dissolving in sterile saline, which also served as the VEH. We utilized three doses of GAL (0.5, 2.0, and 5.0 mg/kg) or VEH (1 mL/kg), which were administered intraperitoneally beginning 24 h after CCI or sham surgery and once daily for 27 days (Experiment 1) or 22 days (Experiment 2). Moreover, treatments were administered 30 min prior to behavioral testing. On the day of microdialysis (Experiment 2), injections were administered after a 3-h baseline collection period. GAL doses and administration route were selected based on published data from our laboratory.^16–19

Experiment 1: the abbreviated 3-choice serial reaction time task

The 3-CSRT task is a modified version of the established 5-CSRT task.^24–26 We previously demonstrated that the 3-CSRT at the 300 ms cue level is sensitive to TBI-induced deficits in male and female adults rats.²⁷ We abbreviated the 3-CSRT to the 2-s cue duration for training efficiency and used it in this study after demonstrating that it is also sensitive to TBI-related chronic attentional deficits.²⁰ We utilized operant chambers (Harvard Apparatus, Holliston, MA) that featured three LED-illuminated cue nose-poke photosensor apertures on one wall, an LED-illuminated food magazine on the opposite wall set to dispense 45 mg sugar pellets (Dustless Precision Pellets^®, Bio-Serv, Flemington, NJ), and a ceiling-mounted house light. Briefly, rats were trained in three successive attentional levels, gradually decreasing the duration of pseudorandomly presented light cues (e.g., 15-s, → 5-s, → 2-s). Correct responses, indicated by a nose-poke in the illuminated aperture, were rewarded with one sucrose pellet. Incorrect responses included nose-pokes to nonilluminated cue holes, whereas trial omissions occurred when rats failed to respond within 5-s post-stimulus presentation. An 8-s intertrial interval (ITI) preceded the next cue presentation. Rats progressed through training levels upon meeting specific performance criteria, achieving >80% accuracy and <20% omissions for three consecutive sessions. Surgery readiness was established when rats achieved criterion performance at the 2-s level, which was considered baseline performance for post-injury behavioral analysis. Dependent measures included percent accuracy for sustained attention [(number of correct responses/total responses) × 100] and percent omissions for distractibility [(number of incorrect responses/total responses) × 100]. Graphic State Notation 4™ software was used to generate experiment protocols and record responses.

Experiment 1: Post-injury 3-CSRT testing

After CCI or sham injury, the rats underwent a 14-day recovery period with mild food restriction reintroduced on day 7 post-injury. Testing on the 3-CSRT task at the 2-s light cue duration began on post-operative day 21 and continued daily for 5 days.

Experiment 2: in vivo microdialysis collection

On post-injury day 21, microdialysis probes (MAB 6.9.2, SciPro, Sanborn, NY) were primed with artificial cerebrospinal fluid (aCSF) [126.5 mM NaCl, 2.4 mM KCl, 1.1 mM CaCl₂, 0.83 mM MgCl₂, 27.4 mM NaHCO₃, and 0.5 mM KH₂PO₄]. In addition, 50 nM neostigmine (Abcam) was added to the aCSF to prevent degradation of ACh levels in the dialysate. Anesthesia was induced as previously described to attach the spring tether and insert a microdialysis probe (3 mm length, 0.6 mm diameter, 15 kD molecular cut off, SciPro, Sanborn, NY) into the implanted guide cannula. The inlet cannula of the probe was connected by a tubing adapter (CMA) and a length of FEP tubing (Harvard Apparatus, 0.12 mm ID) to a low-torque, dual-channel quartz-lined swivel (InStech Inc.), which was serially connected to a syringe mounted on a micro perfusion pump that continuously perfused aCSF + 50 nM neostigmine at a rate of 2 μL/min. The outlet cannula of the probe was connected by a length of FEP tubing to the outlet of the dual-channel swivel and collected into an Eppendorf tube. The swivel was affixed to a multiaxis counter-balanced lever arm (InStech Inc.) attached to a clear polycarbonate enclosure (Cambro, 37.8 cm × 38.1 cm).

Following the probe insertion, aCSF + 50 nM neostigmine was allowed to continuously perfuse for 3 h to allow equilibration prior to the collection of microdialysate for analysis. The rats were then injected with either GAL or VEH, and microdialysate samples were collected every 15 min for a period of 120 min. Samples were immediately placed in dry ice and stored at −80°C until analysis. At the end of the experiment (i.e., after the five post-injection samples were attained), the probe was removed, and the guide cannula obturator was inserted.

Histology: Cortical lesion volume and confirmation of probe placement

Following completion of behavioral testing or microdialysis sampling rats were anesthetized with Fatal-Plus^® (0.3 mL, i.p.) and transcardially perfused with 200 mL of 0.1 M phosphate-buffered saline, followed by 500 mL of 4% paraformaldehyde in 0.1 M phosphate-buffered saline (PBS). The brains were then extracted, fixed in 4% paraformaldehyde, dehydrated with alcohols, and embedded in paraffin wax. Coronal sections (7 μm thick) were cut at 1-mm intervals through the lesion using a rotary microtome and mounted on Superfrost^®/Plus microscope slides. After drying at room temperature (RT), slides were deparaffinized in xylenes, rehydrated, and stained with Cresyl violet to quantify lesion volumes and confirmation of the guide cannula and probe placement. Cortical lesion volumes (mm³) were quantified by first calculating the area of the lesion (mm²), which was achieved by outlining the inferred area of missing tissue for each section (typically 5–7) taken at 1-mm intervals through the extent of the lesion, and then by summing the areas of the lesion, as previously reported.^21,28–30 Sections for each brain in experiment 2 were imaged and magnified (10×), then stitched, and subsequently manually traced and overlaid to determine the probe tract.

Histology: Choline acetyltransferase

Sections were treated with a polyclonal anticholine transferase antibody (Millipore, Temecula, CA) at a concentration of 1:800 in 0.01 M PBS containing 0.1% Triton X-100. The sections were then incubated with the primary antibody for 24 h at RT. After three washes in PBS, the sections were exposed to the secondary antibody (anti-goat IgG; 1:1000; Vector Laboratories, Burlingame, CA) for 1 h at RT. Following incubation with avidin-biotin peroxidase complex (Vector Laboratories) for 1 h at RT and processed using the DAB peroxidase substrate (Vector Laboratories), the reaction was halted by washing the tissue with PBS 3× for 5 min each. Sections were air-dried and cover slipped. Analysis was conducted on a single section underlying the nucleus basalis of Meynert (nbM) area at −0.72 mm from bregma, as per the rat atlas of Paxinos and Watson.²³ ChAT-positive cells in the nbM were quantified using a Nikon i90 microscope (Nikon Corporation, Tokyo, Japan) with a 40× objective.^21,22 Two observers blinded to experimental conditions performed the analysis for each group.

Experiment 2: Measurement of evoked ACh release

The concentrations of ACh were analyzed by HPLC with electrochemical detection (Amuza, San Diego, CA, USA). In this method, a reverse-phase column (2.0 × 150 mm) followed by a post-column immobilized enzyme reactor (Eicom, 1.0 × 4.0 mm) was used. The analyte was subsequently electrochemically detected across a platinum electrode at 450 mV relative to a platinum reference electrode. Chromatograms were analyzed using the Envision software program. ACh peaks were quantified against a 100-fold calibration curve (100 nM, 50 nM, 10 nM, 1 nM). The limit of detection was 10 femtomoles/10 μL injection.

Statistical analyses

Statistical analyses were conducted using StatView 5.0.1 software (Abacus Concepts, Inc., Berkley, CA) by researchers blinded to group conditions. Data acquired over days or multiple sampling were analyzed by repeated measures analysis of variance (rmANOVA). Acute neurological assessments, lesion volume, and immunocytochemical data were analyzed using one-factor ANOVAs. When significant ANOVA main effects or interactions were indicated, the Newman–Keuls post hoc test was used to determine specific group differences. Results are expressed as the mean ± standard error of the mean (SEM), and significance for all analyses was set at p ≤ 0.05.

Results

Experiment 1

There were no significant differences in behavioral outcomes among the sham groups, regardless of treatment (p > 0.05), and thus, the data were pooled into one group designated as SHAM. In Experiment 1, one rat died post-injury, two were removed from each of the TBI + VEH and TBI + GAL (2 mg/kg) groups due to being statistical outliers across several testing days. Therefore, a final total of 63 rats were used in Experiment 1, and the group sizes were as follows: TBI + VEH (1 mL/kg; n = 12), TBI + GAL (0.5 mg/kg; n = 10), TBI + GAL (2.0 mg/kg; n = 13), TBI + GAL (5.0 mg/kg; n = 10), or SHAM (n = 18).

3-CSRT training

All rats reached criterion performance at the testing level (i.e., 2-s cue duration) after an average of 12.7 ± 0.4 training sessions. Across all cue levels, rats achieved criterion performance of >80% accuracy and <20% omission rate. Cue durations of 15-s, 5-s, and 2-s required 3.9 ± 0.2, 3.4 ± 0.1, and 5.2 ± 0.4 training sessions on average, respectively.

Acute neurological assessments

No statistical differences were observed among the TBI groups in hindlimb withdrawal reflex after a brief paw pinch (left range, 151.0 ± 5.6 s to 162.2 ± 6.2 s, p > 0.05; right range, 147.0 ± 5.6 s to 158.5 ± 6.1 s, p > 0.05) or return of righting reflex (range 310.9 ± 18.0 to 353.9 ± 17.2 s, p > 0.05) after the cessation of anesthesia. These data suggest that all TBI rats received an equivalent injury. No reflex differences were revealed among the SHAM groups, and they were significantly shorter than those of the TBI rats, as expected: (left mean = 18.7 ± 1.3 s; right mean = 14.6 ± 1.4 s, p < 0.0001) and righting reflex (113.6 ± 4.7 s, p < 0.0001).

Post-injury 3-CSRT testing: percent accuracy

Before surgery, there were no significant differences in baseline percent accuracy at the 2-s cue duration among the SHAM and TBI groups (F_4,58 = 1.1, p > 0.05), which indicate that all rats reached criterion performance (Fig. 2A). A repeated measures ANOVA revealed that TBI significantly reduced percent accuracy in all injured groups compared with the SHAM controls (Group effect: F_4,58 = 23.2, p < 0.0001; Day effect: F_5,290 = 140.4, p < 0.0001; Group × Day interaction: F_20,290 = 8.6, p < 0.0001). The Newman–Keuls post hoc analysis revealed that the 0.5 mg/kg and 2.0 mg/kg GAL doses neither improved TBI-induced reductions in percent accuracy nor reduced percent accuracy (p > 0.05). These findings contrast with the 5.0 mg/kg GAL dose that further reduced percent accuracy relative to the GAL (0.5 mg/kg and 2.0 mg/kg) and VEH-treated TBI groups (p < 0.05).

FIG. 2.

Mean (±SEM) % accuracy (A) and % omissions (B) averaged for the last three baselines at the 2 s cue and then each day of 3-CSRT testing post-surgery. Repeated measures ANOVA revealed no statistical differences among the sham groups (n = 3–5/group), and therefore, they were pooled and designated as SHAMs (n = 18). (A) Three weeks following injury, rats displayed significant and persistent impairments in % accuracy, a measure of sustained attention. Student–Newman–Keuls post hoc tests revealed that the low (0.5 mg/kg) and moderate (2.0 mg/kg) galantamine (GAL) doses did not attenuate traumatic brain injury (TBI)-induced attentional deficits relative to vehicle (VEH) and that the higher dose (5.0 mg/kg) exacerbated impairments. (B) Assessment of % omissions, a measure of distractibility, indicated that GAL (0.5, 2.0, and 5.0 mg/kg) did not restore TBI-induced deficits. ^#p < 0.05 for SHAM vs. TBI + VEH and TBI + GAL (0.5, 2.0, and 5.0 mg/kg). *p < 0.05 for TBI + GAL (5.0 mg/kg) vs. TBI + VEH and TBI + GAL (0.5 and 2.0 mg/kg). n = 10–18/group.

Post-injury 3-CSRT testing: percent omissions

There were no significant differences in baseline percent omissions at the 2-s cue duration among the SHAM and TBI groups (F_4,58 = 4.7, p > 0.05), which indicate that all rats reached criterion performance (Fig. 2B). A repeated measures ANOVA showed that TBI significantly increased percent omissions in all injured groups compared with the SHAM controls (Group effect: F_4,58 = 5.5, p < 0.01; Day effect: F_5,290 = 58.0, p < 0.0001; Group × Day interaction: F_20,290 = 3.8, p < 0.0001). Newman–Keuls post hoc analysis revealed no differences between the VEH- and GAL-treated TBI groups, regardless of GAL dose.

Post-injury 3-CSRT testing: individual cue port percent accuracy

Left cue port

There were no significant differences in baseline left cue port percent accuracy at the 2-s cue duration between the SHAM and TBI groups (F_4,58 = 34.4, p > 0.05), indicating that all groups performed equally prior to surgery (Fig. 3A). Repeated measures ANOVA revealed that TBI significantly reduced left-cue port percent accuracy in all injured groups compared with the SHAM controls (Group effect: F_4,58 = 28.3, p < 0.0001; Day effect: F_5,290 = 104.0, p < 0.0001; Group × Day interaction: F_20,290 = 6.8, p < 0.0001). TBI-induced deficits were significantly exacerbated by the 5.0 mg/kg GAL dose compared with the 2.0 mg/kg dose and VEH. There were no differences among the 0.5 mg/kg or 2.0 mg/kg doses of GAL and VEH-treated TBI groups (p > 0.05).

FIG. 3.

Mean (±SEM) % accuracy in the left (A), center (B), and right (C) cue port holes averaged for the last three baselines (BSL) at the 2 s cue and then each day of 3-CSRT testing post-surgery. Statistical analysis of BSL performance revealed that there were no significant differences in % accuracy across the cue ports and among groups prior to surgery indicating a lack of spatial bias. (A) Following TBI, pronounced % accuracy deficits to the left cue port are suggestive of hemi-spatial neglect of the left visual field contralateral to the injury site. GAL did not restore TBI-induced deficits of sustained attention, relative to VEH, and the higher dose of GAL (5.0 mg/kg) exacerbated attentional impairments. (B) Assessment of the center % accuracy indicates that post-injury GAL administration regardless of dose was associated with inferior attentional performance. (C) TBI induced significant attentional impairments to the right cue port, which were exacerbated by TBI + GAL (5.0 mg/kg). ^#p < 0.05, SHAM vs. TBI + VEH and TBI + GAL (0.5, 2.0, and 5.0 mg/kg). *p < 0.05, TBI + GAL (5.0 mg/kg) vs. TBI + VEH and TBI + GAL (0.5 and 2.0 mg/kg). ^p < 0.05, TBI + VEH vs. TBI + GAL (0.5, 2.0., and 5.0 mg/kg). n = 10–18/group.

Center cue port

There were no significant differences in baseline center-cue port percent accuracy at the 2-s cue duration between the SHAM and TBI groups (F_4,58 = 12.7, p > 0.05), which indicate that all groups performed equally prior to surgery (Fig. 3B). Repeated measures ANOVA revealed that TBI significantly reduced percent accuracy in the center-cue port in all injured groups compared with the SHAM controls (Group effect: F_4,58 = 11.2, p < 0.0001; Day effect: F_5,290 = 65.4, p < 0.0001; Group × Day interaction: F_20,290 = 4.5, p < 0.0001). Post-injury administration of GAL, regardless of dose, exhibited significant reductions in the center-cue port compared to the VEH-treated TBI group (p < 0.05).

Right cue port

There were no significant differences in baseline right-cue port percent accuracy at the 2-s cue duration between the SHAM and TBI groups (F_4,58 = 21.4, p > 0.05), which indicate that all groups performed equally prior to surgery (Fig. 3C). Repeated measures ANOVA revealed that TBI induced a significant reduction in the right cue port percent accuracy in all injured groups compared to the SHAM controls (Group effect: F_4,58 = 11.4, p < 0.0001; Day effect: F_5,290 = 35.3, p < 0.0001; Group × Day interaction: F_20,290 = 3.5, p < 0.0001). Within the TBI groups, the 5.0 mg/kg GAL dose significantly reduced accuracy compared with the 2.0 mg/kg GAL dose and VEH (p < 0.05).

Lesion volumes

One-way ANOVA revealed no differences (F_3,34 = 2.1, p > 0.05) in cortical lesion volumes (mm³) among the GAL 0.5 mg/kg (56.1 ± 4.3 mm³), 2.0 mg/kg (42.5 ± 4.9 mm³), 5.0 mg/kg (56.1 ± 4.3 mm³), and VEH-treated (47.2 ± 3.9 mm³) groups at four weeks post-injury (Fig. 4A). Figure 4B depicts mean representative sections of the 4 TBI groups as follows: VEH, upper left; GAL (0.5 mg/kg), upper right; GAL (2.0 mg.kg), lower left; and GAL (5.0 mg/kg), lower right.

FIG. 4.

Mean (±SEM) cortical lesion volume (mm³) at 4 weeks post-TBI. (A) There were no statistical differences (p > 0.05) in lesion volumes among groups suggesting that galantamine (GAL) did not attenuate TBI-induced cortical tissue loss regardless of dose, relative to vehicle (VEH). (B) Bright field image (stitched) (10×) of Cresyl violet-stained representative coronal sections (7-μm thick, Bregma ∼ −1.20 mm). Upper left = TBI + VEH; Upper right = TBI + GAL (0.5 mg/kg); Bottom left = TBI + GAL (2.0 mg/kg); Bottom right = TBI + GAL (5.0 mg/kg). n = 8–11/group.

ChAT+ cells in the right and left nbM

One-way ANOVAs revealed no group differences in ChAT+ cells in the left nbM (F_4,30 = 1.775, p > 0.05; Fig. 5A), but a significant difference was revealed in the right nbM (F_4,30 = 4.9, p < 0.05; Fig. 5B). Post hoc analyses revealed that the SHAM control group had significantly more ChAT+ cells in the ipsilateral nbM relative to the GAL (0.5, 2.0, 5.0 mg/kg) and VEH-treated TBI groups. Figure 5C depicts ChAT+ cells sampled from the left and right nBM. Figure 5D depicts mean representative sections of the SHAM and TBI groups.

FIG. 5.

Mean (±SEM) choline acetyl-transferase positive (ChAT+) cells in the nucleus basalis of Meynert (nbM). (A) ChAT+ cells in the left nbM (contralateral to CCI). There were no statistical differences in the number of cells across groups (p > 0.05). (B) ChAT+ cells in the right nbM (ipsilateral to CCI). TBI significantly decreased ChAT+ cells in the right nbM, regardless of treatment condition (p < 0.05). (C) Bright field image (stitched) (10×) of a ChAT-stained coronal section following CCI (7-μm thick, Bregma −1.20 mm). Red boxes in the left and right hemisphere indicate the nbM. (D) Representative images of ChAT+ cells in the left and right nbM. ^#p < 0.05, SHAM vs. TBI + VEH and TBI + GAL (0.5, 2.0, and 5.0 mg/kg). n = 5–9/group.

Experiment 2

One rat from the TBI + GAL (0.5 mg/kg) group died post-injury. An additional rat from the TBI + GAL (0.5 mg/kg) and two from the TBI + VEH group were omitted from the analyses because they were below the lower limit of quantification.³¹ Hence, the final groups for ACh dialysate collection were TBI + VEH (1 mL/kg; n = 2), TBI + GAL (0.5 mg/kg; n = 3), TBI + GAL (2.0 mg/kg; n = 4), and TBI + GAL (5.0 mg/kg; n = 4).

Acute neurological assessment

During phase 1 (CCI), no statistical differences were observed among the TBI groups in hindlimb withdrawal reflex (left range, 153.5.0 ± 7.8 s to 161.0 ± 9.6 s, p > 0.05, right range, 143.00 ± 9.2 s to 156.2 ± 9.7 s, p > 0.05) or return of righting reflex (range 334.75 ± 56.5 s to 417.6 ± 31.2 s, p > 0.05). In phase 2 (guide cannula implant), there were no statistical differences observed after a brief paw pinch following the cessation of anesthesia (left range, 18.3 ± 5.7 s to 54.5 ± 25.7 s, p > 0.05, right range, 12.6 ± 4.8 s to 40.2 ± 23.3 s, p > 0.05).

In vivo microdialysis in freely-moving rats

The repeated measures ANOVA revealed a significant Group × Sample interaction (F_7,21 = 3.149, p < 0.01). The Newman–Keuls post hoc revealed that prior to injection, there were no differences in basal ACh efflux regardless of VEH or GAL treatment in injured rats (p > 0.05; Fig. 6A). Following injection, the 5.0 mg/kg GAL TBI group exhibited significantly increased ACh efflux in the mPFC compared with the 0.5 mg/kg and 2.0 mg/kg GAL and VEH-treated TBI groups (p < 0.05), which did not differ from one another (p > 0.05; Fig. 6A). Chromatogram depicted the mean peak of ACh for the 4 TBI groups (Fig. 6B).

FIG. 6.

(A) Changes in acetylcholine (ACh) levels in the right mPFC of injured rats prior to and following GAL or VEH treatment. Injections were administered after a 45-min baseline collection period (B1, B2, B3). Samples were collected at 15-min intervals after injection (S1, S2, S3, S4, S5). The data are expressed as a percentage of the mean basal efflux (±SEM). (B) Overlaid representative chromatograms obtained from a 20 µL injection of dialysate S2 (i.e., 30-min post-injection) from TBI + VEH, TBI + GAL (0.5 mg/kg), TBI + GAL (2.0), and TBI + GAL (5.0 mg/kg). The analyte was electrochemically detected across a platinum electrode at 450 mV relative to a platinum reference electrode and recorded as a function of millivolts (mV) to time (min). The detector response is directly proportional to concentration (1.8–2.0 mV $\approx$ 100 fmol). ACh concentration was quantified against a 100-fold calibration curve of ACh perchlorate standard with a retention time of 12.7 min. The colored lines correspond to drug-treated groups: gray, TBI + VEH; blue, TBI + GAL (0.5 mg/kg); green, TBI + GAL (2.0 mg/kg); and orange, TBI + GAL (5.0 mg/kg). *p < 0.05 for TBI + GAL (5.0 mg/kg) versus TBI + VEH and TBI + GAL (0.5 and 2.0 mg/kg). n = 2–4/group.

Lesion volumes

One-way ANOVA revealed no group differences (F_3,11 = 0.4522, p > 0.05) in cortical lesion volumes (mm³) among the GAL 0.5 mg/kg (32.0 ± 7.9 mm³), 2.0 mg/kg (35.2 ± 4.4 mm³), 5.0 mg/kg (39.1 ± 4.8 mm³), and VEH-treated (46.7 ± 14.8 mm³) TBI groups (Fig. 7A). Figure 7B is a digital rendering of the overlapping probe tracks in the right mPFC from each rat. Figure 7C is a representative image of the guide cannula (i.e., indicated by the bracket), and the asterisk indicates the deepest point of the probe.

FIG. 7.

(A) Mean (±SEM) cortical lesion volume (mm³) at 4 weeks post-TBI. There were no statistical differences (p > 0.05) in lesion volumes among groups suggesting that GAL did not attenuate TBI-induced cortical tissue loss regardless of dose, relative to VEH. TBI + VEH, n = 4; TBI + GAL (0.5 mg/kg), n = 3; TBI + GAL (2.0 mg/kg), n = 4, TBI + GAL (5.0 mg/kg), n = 4. (B) Bright field image (stitched) (10×) of Cresyl violet-stained coronal sections (7-µm thick, Bregma ∼ −1.20 mm). (C) Digital rendering of the probe tracks in the right mPFC for each rat in experiment 2. The colored lines correspond to drug-treated groups: gray, TBI + VEH; blue, TBI + GAL (0.5 mg/kg); green, TBI + GAL (2.0 mg/kg); and orange, TBI + GAL (5.0 mg/kg). (D) A guide cannula was stereotaxically placed using an 8-degree lateral approach at coordinates AP +2.6 mm, ML: +1.1 mm from bregma. On the day of dialysate collection, a probe was inserted into the guide cannula with the membrane extending 3 mm beyond the guide cannula in the mPFC. The asterisk indicates the deepest point of the probe.

Discussion

The aim of this study was to evaluate the effectiveness of GAL, a dual-action competitive AChE inhibitor and nicotinic acetylcholine receptor (nAChR) positive allosteric modulator, provided chronically after injury, on sustained attention performance and ACh efflux from the mPFC. We demonstrated that contrary to our hypothesis, GAL administration did not remediate attentional impairments, cortical cavitation, or injury-induced reductions of cholinergic neurons in the ipsilateral nbM regardless of given doses. Notably, the higher dose (5.0 mg/kg) of GAL exacerbated sustained attention deficits. In vivo sampling from mPFC revealed no differences in basal ACh levels; however, the higher dose of GAL significantly evoked ACh release, peaking at 30 min post-administration in injured rats, which aligns with the start of 3-CSRT testing. While the higher GAL dose increase evoked ACh release beyond basal levels in the mPFC, it did not ameliorate attentional deficits. Together these findings support a dose-dependent effect of GAL on attentional performance and drug-evoked release of ACh in the injured brain suggesting that transiently increased cholinergic tone does not necessarily render attentional and histological improvements.

Although complex behavioral tasks like the 3-CSRT are well-established, they are relatively novel in pre-clinical TBI.²⁷ Therefore, it is essential to highlight behavioral measures and corresponding behavioral correlates to identify distinct aspects of attention impacted by injury and pharmacological manipulations. Briefly, a correct response (i.e., poking into the illuminated cue port) suggests intact attentional control. Conversely, failure to respond within the designated stimulus duration can be due to detection failures, repeats of responses, or failure to suppress ongoing behaviors (e.g., rearing or grooming) and, collectively, is a behavioral correlate of distractibility. Together, impaired or weakened attentional control contributes, in part, to increased distractibility. Additional factors, such as injury and pharmacological manipulations along with external distractors, impose greater demands on top-down attentional control and can contribute to impaired sustained attention. Our laboratory has previously demonstrated the sensitivity of the 3-CSRT to injury alone on TBI-induced attentional deficits in adult male and female rats.²⁷ In this study, we demonstrate a dose-dependent impact of a pharmacological intervention in combination with injury on attentional performance in the 3-CSRT. This effect is particularly evident in the injury group that received the higher dose of GAL, demonstrating exacerbated impairments in sustained attention and cue-port specific accuracy compared with the low and moderate GAL and VEH-treated TBI groups. Conversely, behavioral measures of distractibility were not susceptible to the combined detrimental effects of the higher GAL dose after TBI. In this context, exacerbated impairments to sustained attention without greater deficits in distractibility suggest a drug-mediated impact on the ability to maintain consistent focus and accuracy to respond to the stimulus (i.e., sustained attention) in the injured brain, as opposed to increased susceptibility from nontarget stimuli or irrelevant environmental factors (i.e., distractibility). In addition, in the 3-CSRT task, the rats are required to maintain scanning attention across the three spatially distinct ports revealing cue-port specific measures of sustained attention. Accuracy impairments were most pronounced when the cue was presented in the left-side hole, mirroring clinical unilateral spatial neglect of the visual field contralateral to the site of injury, which we previously also reported using the 300 ms cue duration.²⁷ In addition, the higher dose of GAL exacerbated TBI-induced deficits to the left and right cue port compared to the low and moderate GAL- and VEH-treated injury groups. Interestingly, GAL treatment, regardless of dose, impaired center cue-port accuracy, suggesting a drug-mediated effect on visual attentional scanning or cue detection compared to the VEH group. These distinct attentional measures are suggested to be underpinned by cholinergic tone and receptor-dependent mechanisms.

Several seminal studies have highlighted the importance of transient cholinergic signaling during tasks that demand sustained attention, which in combination with TBI studies demonstrating chronic downregulation of the cholinergic system after injury, provided the rationale necessary to examine the efficacy of a pro-cholinergic pharmacotherapy in this study. Early microdialysis studies in non-TBI rats demonstrated an increase in ACh levels in the mPFC during the 5-CSRT task.^32,33 In addition, targeted immunotoxic lesions affecting BF cholinergic neurons led to impaired sustained attention and reduced task-evoked ACh release in the mPFC.³⁴ These findings emphasize the critical role of intact cholinergic neurons originating in the BF in regulating attentional performance, which can be particularly susceptible to TBI as demonstrated by our results herein. Although contrary to our hypothesis, our results demonstrated that 27 days after TBI, significant deficits in ChAT+ cells in the ipsilateral BF remained unchanged despite chronic GAL treatment. It is possible that injury-induced disruptions to BF cholinergic neurons, which send dense cholinergic projections to the mPFC,³⁴ may impair ACh signaling, thereby contributing, in part, to reduced attentional performance. Beyond intact cholinergic neurons, non-TBI studies also examined the influence of the alpha (α) and beta (β) subunits of nAChRs on distinct attentional measures (e.g., sustained attention and distractibility). Knockout of β2 subunits in the prelimbic cortex of mice resulted in more omission errors, whereas reexpression of these subunits alleviated the deficits.³⁵ Conversely, mice lacking the α5 subunit exhibited decreased accuracy without affecting omissions significantly. Moreover, pharmacological studies demonstrated that nicotine, an agonist of these receptors, improves sustained attention, but this effect is lost in mice lacking α7 nAChRs.^36,37 After TBI, there are reductions in the α3, α4, and α7 nAChR subunits.^11,12,38 Reduced α7 receptor binding was observed after CCI, affecting regions such as the hippocampus and somatosensory cortex, with changes detected within hours to days post-injury,³⁸ whereas reductions in α3 and α4 subunits were delayed, occurring days to weeks after the injury.³⁸ Therefore, impaired attention may be due to a combined injury-induced reduction in cholinergic tone and subunit specific nAChR expression. Moreover, the pathogenic effect of the higher GAL dose on accuracy, but not omissions, may be due to a dose-dependent effect of high GAL on the desensitization of the α subunit of nAChRs previously implicated in accuracy as opposed to the β nAChR subunits, which demonstrated a role in omissions.

The lower and moderate doses of GAL did not increase in vivo ACh release in the mPFC after injury. Dixon and colleagues showed significant reductions in drug evoked ACh release in the hippocampus³⁹ and neocortex⁴⁰ after TBI. Therefore, it is plausible that albeit low and moderate doses of GAL enhanced nAChR sensitivity, TBI-induced reductions of ACh concomitant with limited drug-evoked ACh release neither improved nor exacerbated attentional functioning or ChAT+ in the ipsilateral BF after TBI. Conversely, results from Experiment 2 revealed that the higher dose of GAL significantly increased in vivo ACh efflux in the mPFC. We administered GAL daily as an adjuvant therapy, with the drug given 30 min before the onset of the 3-CSRT task to align with the peak penetration of GAL in the brain.⁴¹ Despite drug-evoked ACh efflux, the high GAL dose exacerbated TBI-induced deficits in sustained attention, suggesting that there is an optimal dose of ACh modulation for attentional functioning. To this end, the high GAL dose did not attenuate TBI-induced ChAT cell loss, which may be reflective of endogenous downregulation of the ChAT enzyme to produce ACh.

In cases of pathologically low cholinergic activity such as Alzheimer’s disease or TBI, stimulating ACh can enhance attention, learning, and memory. Conversely, excessive ACh signaling can disrupt neuronal activity, including cholinergic receptor desensitization. Therefore, it is possible that the high dose of GAL exacerbated TBI-induced deficits because of increased AChE inhibition, like what is observed following high doses of donepezil after TBI. Shaw and colleagues demonstrated that higher doses of donepezil, a selective AChE inhibitor, exacerbated deficits in learning and memory after TBI.¹⁴ Furthermore, artificially increasing ACh levels by physostigmine, an AChE inhibitor, impairs memory consolidation and retrieval in rodent and human subjects.^42–44 GAL has two separate mechanisms of action to enhance cholinergic neurotransmission.⁴¹ At low doses, GAL binds allosterically to the α7 nAChRa, increasing ACh efflux from presynaptic neurons. At high doses, GAL functions more closely as an AChE inhibitor. Therefore, the “high” GAL dose in this study (5 mg/kg) has the potential to preferentially inhibit ACh degradation while maintaining the potential to enhance ACh efflux through allosteric receptor potentiation. Although not directly examined in this study, it is possible that high ACh input due to AChE inhibition may activate autoinhibitory metabotropic muscarinic AChRs and thereby suppress the magnitude of feedback excitation in the cortex through presynaptic inhibition of glutamate release. Alternatively, extended exposure to ACh can cause prolonged desensitization of nAChRs, limiting the receptors’ response to endogenous ACh. Therefore, alternative pro-cholinergic drugs such as type I and type II α7 nAChR positive allosteric modulators (PAMs) may enhance receptor function without disrupting or slowing the desensitization kinetics of the receptor, respectively. Recent studies from our laboratory^27,45 and others support the therapeutic potential of α7 nAChR PAMs on restoring measures of attention and contextual fear conditioning after TBI.^46,47

Previous studies from our laboratory examined the therapeutic efficacy of GAL on complementary neurobehaviors after TBI. Specifically, de la Tremblaye and colleagues examined three chronic doses of GAL (1.0, 2.0, and 3.0 mg/kg) and revealed that only the 2 mg/kg daily administration paradigm improved learning and memory, as assessed in a Morris water maze task in adult male rats after TBI.¹⁷ Interestingly, the lack of behavioral benefit extended by the high GAL dose underscores similar findings from the current study, suggesting that an optimal level of drug evoked ACh is necessary to mediate therapeutic efficacy. Furthermore, regardless of these doses, GAL similarly did not mitigate cortical cavitation.¹⁷ We also demonstrated that chronic GAL at 1.0 and 2.0 mg/kg/day improved cognitive flexibility, as assessed by the attentional set-shifting task, with the latter dose reducing cortical cavitation after TBI.¹⁹ Cognitive flexibility, the ability to adaptively react to changing stimuli from the environment, is highly modulated by ACh signaling in the dorsal-medial striatum, hippocampus, basolateral amygdala, and posterior parietal cortex.⁴⁸ While several studies suggest that ACh in the PFC plays a crucial role in modulating cognitive flexibility, the precise contribution of cholinergic signaling in this function remains unclear. Some reports indicate that ablation of cholinergic neurons from the rat nbM does not impair reversal learning.^34,49 Therefore, it is possible that GAL-mediated benefits in the hippocampus and dorsal striatum rescued TBI-induced deficits in cholinergic signaling in brain regions relevant to hippocampal-dependent learning and memory and cognitive flexibility, but not that of sustained attention, which relies substantially on the PFC. Alternatively, distinct roles for the monoaminergic system, including dopamine (DA) and norepinephrine (NE) in visual attention have been demonstrated in comparison to the cholinergic system during the performance of the 5-CSRTT.^33,50 These studies reported separate functions for NE and ACh in attention and cognitive processes such that efflux of NE was observed when task contingencies of sustained attention are extinguished.³³ Conversely, sustained ACh efflux was observed only during performance of the task,³³ whereas DA regulation may contribute to successful response selection in mPFC-based attentional tasks.⁵⁰ Therefore, catecholamine agents like methylphenidate (MPH), a FDA approved pharmacotherapy for treating attention deficit hyperactivity disorder, thought to act as a DA agonist to increase DA and NE in the PFC may also extend efficacy on the abbreviated 3-CSRT following TBI. In non-TBI studies examining the efficacy of MPH treatment on 5-CSRTT attentional performance has exhibited mixed results demonstrating no effect, improvements to accuracy (i.e., sustained attention), or increased measures of impulsivity.^51–54 These discrepancies may be due to varying paradigms methodologies, including variable task ITI.^51–54 However, to date, there are no published studies documenting beneficial effects of MPH on task of complex cognition after pre-clinical TBI. Pre-clinical assessments of pharmacotherapies targeted to enhance DA signaling have shown mixed results after TBI, notably amantadine and D-amphetamine, such that a high dose of D-amphetamine significantly reduced impulsivity for only severely injured rats in the 5-CSRTT, whereas amantadine-induced reductions in motor impulsivity were accompanied by motivational slowing across groups.⁵⁵

A potential limitation of this study includes the limited sample size and lack of sham controls during microdialysis procedures. The former is partly due to experimental constraints, including injury-induced death and sample elimination due to dialysate samples falling below the limit of quantification according to the recommended guidelines.³¹ Previous studies have demonstrated no differences in basal ACh levels in the hippocampus and neocortex between sham and TBI rats without a pharmacological challenge.^39,40

Conclusion

This study reveals that chronic GAL across three clinically relevant doses (0.5, 2.0, and 5.0 mg/kg) was not efficacious at restoring sustained attention or distractibility after TBI, despite the higher dose increasing cholinergic tone as assessed by in vivo microdialysis. Using behavioral assessments such as the 3-CSRT, which are analogous to clinical measures of attentional functioning, in combination with direct examination of the drug-evoked response on ACh concentration in the brain, provides greater insight into the therapeutic and translational efficacy of drug treatments for TBI. The combined assessment of behavior and the transient dynamics of GAL on in vivo ACh release over time in specific brain regions, particularly the mPFC, a behaviorally-relevant brain region, further elucidates the regional effects of GAL on cholinergic signaling in the injured brain.

Transparency, Rigor, and Reproducibility Summary

The study was not preregistered in a public domain but was prespecified based on prior work from our group with similar study and statistical plans. Sample sizes were calculated to have >80% power to detect significant effects after correcting for multiple comparisons. Prior to surgical manipulations, the rats were randomly assigned to group conditions. Investigators performing behavioral assessments and histological analyses were blinded to surgical (TBI vs. Sham) and treatment (GAL vs. VEH) conditions. All behavioral assessments were conducted between 07:00 and 19:00 to remain consistent with previous works. Statistical measures included ANOVAs, and when appropriate the Newman–Keuls post hoc test was used to determine specific group differences. Results are expressed as the mean ± standard error of the mean (SEM), and significance was set at p ≤ 0.05. All procedures were performed in accordance with the corresponding author’s university IACUC. Data from these experiments will be available in a FAIR compliant data repository (odc-sci.org). All behavioral testing equipment, therapeutics, software, and microdialysis probes are commercially available.

Authors’ Contributions

E.H.M., A.E.K., and C.O.B. developed the study. E.H.M., A.E.K., and C.O.B. wrote and edited the abstract. E.H.M., H.E.C., E.M.A., V.B.D., J.A.S., H.M.D., N.R.G., P.L.R., R.A.B., and V.J.V. performed the behavioral assessments. E.H.M., A.E.K., C.O.B., and J.P.C. performed surgical procedures. E.H.M. performed microdialysis collection. E.H.M., H.E.C., E.M.A., V.B.D., H.M.D., and P.L.R. performed brain tissue histological assessments. E.H.M. performed dialysate chromatography assessments. E.H.M., A.E.K., and C.O.B. performed statistical analyses. E.H.M., A.E.K., and C.O.B. wrote the manuscript. All authors reviewed and approved the manuscript.

Author Disclosure Statement

No competing financial interests exist.

Funding Information

This work was supported, in part, by the National Institute of Health grants NS110609 (C.O.B.), NS084967, NS121037 (A.E.K.) and the UPMC Children’s Research Advisor Committee Dissertation Fellowship (E.H.M.) and Brain Injury Association of America Dissertation Grant (E.H.M.).