Electrical stimulation methods and protocols for the treatment of traumatic brain injury: a critical review of preclinical research

The excitability of neuronal cells facilitates modulation of their firing activity using external stimulation to enhance or suppress endogenous activity [30]. This modulation can be utilized for therapeutic or diagnostic purposes in several neurological diseases or injuries to nervous tissue [31‐34]. To better understand the advantages of therapeutic electrical stimulation following TBI, it is necessary to gain extensive insights into how and to which extent stimulation influences neuronal physiology and morphology.

Artificial electrical stimulation may change the electrical potential of the surrounding extracellular region through anodic as well as cathodic protocols [35‐37]. In cathodic stimulation, a negative current pulse is delivered to the extracellular area, which in turn depolarizes the cellular membrane with the aim to elicit an action potential. Anodic stimulation instead hyperpolarizes the region near the site of interest and thus decreases the membrane potential [38]. This results in a flux of positive ions towards the stimulation site from surrounding areas, which leads to a depolarization of the cell membrane further away from the site of stimulation, possibly triggering an action potential at the nearest Ranvier node [39, 40].

The effect of electrical stimulation on the brain depends on the inherent characteristics of the tissue. At the cellular level, it is easier to excite an axon than a soma, and myelinated axons are the most excitable part of the cell [41, 42]. Induced voltages differ between nodes and internodes due to the drastic differences in voltage-gated ion channel density [43, 44]. Activated axons progress the signal antidromically to the soma [45, 46] and orthodromically to the synaptic terminals [47, 48]. Bending, branching and significant changes in the diameter of an axon determine the effective site and threshold of the stimulation [43]. Generally, it is easier to elicit action potentials with negative currents in almost all cell compartments, except for some types of dendrites that are more prone to stimulation with positive currents [43, 49, 50].

The most prevalent electrical stimulation methods used in post-TBI treatment studies, which are in the focus of this review, are transcranial magnetic stimulation (TMS), transcranial direct current stimulation (tDCS), deep brain stimulation (DBS) and vagus nerve stimulation (VNS) [25, 79]. TMS and tDCS are amongst the most commonly used non-invasive brain stimulation techniques [80, 81]. They are effective in the treatment of a wide variety of neurologic impairments, but their efficiency and precision is limited by the distance of the stimulator to the target region. Invasive stimulation methods, such as DBS or VNS, may achieve higher precision and efficiency by bringing the stimulation electrodes closer to the desired area. A schematic overview of these four stimulation methods and their preclinical usage is depicted in Fig. 1.

To cover the most commonly used variations that describe the stimulation methods selected for this review as well as TBI, the search query consisted of the following MeSH terms:

(“transcranial magnetic stimulation” OR “transcranial direct current stimulation” OR “deep brain stimulation” OR “vagus nerve stimulation” OR "vagal nerve stimulation") AND (“traumatic brain injury” OR “tbi” OR “concussion”)

The search was conducted in the PubMed and Web of Science databases. To obtain as many relevant records as possible, the query was searched in all fields of the respective databases, which includes titles, abstracts and keywords of publications, among other information. The list of search results was last updated on the 8th of September 2022 and the search yielded 358 results in PubMed and 524 in Web of Science, amounting to a total of 583 different records after removing duplicates. The results were sorted by publication date from oldest to most recent and the titles, authors and publication years of these records were exported from the respective databases and collected in a Microsoft Excel spreadsheet to organize further screening.

The abstracts of all 583 individual search results were screened by one investigator for five different criteria of interest to this review. This was done manually without the use of any advanced automation tools except for a simple text search function. First, the abstract needed to mention TBI as the underlying cause of the disability under investigation. Next, an electrical stimulation method had to be utilized in the study and third it had to be used for a therapeutic purpose or as a treatment, as opposed to a diagnostic application. The record also needed to consist of original research, which excluded other review articles and excerpts from larger studies, such as meeting abstracts and conference papers. Finally, only preclinical studies were included, where an animal model was utilized to investigate certain parameters of interest.

These five criteria were assessed in the order described above, and when an article did not contain that criterion, it was immediately excluded from the review. A total of 543 records were excluded, 93 of which did not investigate TBI sequelae, 60 were not about electrical stimulation, 97 used these methods for an application other than therapy, 162 were not original research and 131 of the remaining articles were not preclinical studies. This abstract screening resulted in 40 articles for the following full-text assessment step.

Out of the 40 articles selected for full-text assessment, another six were excluded. Four of the excluded articles used electrical stimulation not for the therapy but for the assessment of stimulation effects on healthy animals. One study was not original research, which was not immediately apparent in the abstract, and another article did not utilize electrical stimulation altogether. Ultimately, literature screening led to a total of 34 articles that were reviewed in this study. Eight separate studies used TMS and VNS respectively, seven employed tDCS, and eleven utilized DBS for the treatment of TBI sequelae.

Most of the included TMS studies listed in Table 1 investigated the loss of motor functions after TBI [138, 152‐155], while some also used it as a potential treatment for detrimental changes in brain metabolism [152, 156], behavioral impairments [157], and to prevent cell death [152, 156]. A recent study also investigated the mechanisms of rTMS treatment without considering any specific disability [158]. Animals were usually immobilized and awake during stimulation, except for two studies, where TMS was applied during temporary anesthesia using volatile anesthetics [153, 154]. In four of the studies stimulation was done at the ipsilateral side [138, 152, 157, 158], in one at the contralateral side [153], and in two at the medial alignment to the injury site [154, 156]. Stimulus intensities are rarely given as absolute values, such as a magnetic field strength, but as a percentage of the maximum output of the stimulator [138, 153, 156, 157] or of the experimentally determined resting motor threshold of the animal [152, 158], while two studies do not specify the intensity of the stimulation [154, 155]. Many protocols employed rTMS in the form of pulse trains at frequencies ranging from 2 to 40 Hz, some of them having applied the stimulus for 9 to 20 min [138, 153, 155‐158], while others stimulated for 3 min or less [152, 154]. In half of the studies TMS treatment was started 1 day after injury [138, 154, 156, 157], and the other half started stimulation several days later [152, 153, 155, 158]. Stimulation sessions were usually administered daily and continued for 1 week or longer. The target of TMS was often a nonspecific area of the cortex, apart from one study where the primary sensory region on the non-injured side of the brain was stimulated in pediatric animals [153] and another study that specifically targeted subcortical areas [155]. Persisting effects of TMS were rarely investigated, but one research group claims to have found a long-lasting increase of excitability in the non-injured cortex [153], while another found functional improvements lasting for up to 6 weeks after stimulation when TMS was combined with environmental enrichment [154]. Three studies observed a neuroprotective effect and the prevention of cell death [152, 155, 156], while two each determined that TMS could be an effective treatment to improve motor function [138, 155], induce neural plasticity [153, 158], or help with the recovery of brain activity [138, 152]. It was also shown that TMS led to histologic improvements after TBI, meaning that the expression levels of relevant proteins changed towards a positive outcome [155, 157, 158]. Individual studies determined that TMS could decrease hyperactivity [153], improve cell metabolism and at the same time induce cell proliferation and neurogenesis [156], help with the recovery from behavioral impairments [157], improve cortical excitability [154], or enhance cognitive function [155]. Only one study did not observe any improvements in motor function after applying TMS [152].

The studies shown in Table 2 used tDCS mainly to assess improvements in motor function, excitability and cognitive impairments [159‐163], but also its effects on cerebral blood flow (CBF) and tissue oxygenation after TBI [161, 164]. Only one study examined tDCS as a treatment for psychiatric-like symptoms such as impulsivity and attention [165]. Animals were anesthetized during tDCS in four of the seven studies [159, 160, 163, 165] and stimulation was applied for 10–30 min in all studies. In four studies, sessions were repeated for several days and lasted up to 4 weeks [159, 161, 162, 165], while three studies applied the stimulation only once in either the acute [160], subacute [163] or chronic phase [164] after TBI respectively. In six of the seven studies [159‐164] anodal tDCS with an amplitude between 0.1 and 0.2 mA was applied. Nevertheless, the surface area of the employed electrodes varied considerably, resulting in widely different current densities between 0.255 and 2.82 mA/cm², which is a critical factor for effective stimulation [104]. The anode was usually placed near the lesion or motor cortex, and the cathode at the thorax or trunk of the animal. Only one study [165] employed cathodal instead of anodal tDCS with a higher amplitude of 0.8 mA, resulting in a current density of 0.708 mA/cm², whereby the cathode was placed near the bregma and the anode between the scapulae. One group observed a persisting increase in local cortical CBF in response to tDCS in TBI and control animals, as well as improved motor and cognitive outcome 1 week after the end of the stimulation in one of the stimulation groups [161]. However, all other studies in this scope that investigated long-term changes after stimulation [159, 160, 165] found that the beneficial effects of the treatment were no longer apparent after longer observation periods, over which non-treated animals reached a similar level of recovery.

With the possibility to target small and specific areas as well as deeper regions of the brain, DBS can be used to treat a wide variety of different impairments, such as the loss of cognitive [166‐170] and motor function [171], as well as bladder dysfunction [172‐174] and disorders of consciousness [175, 176]. Whilst the stimulation protocols differ greatly with respect to the targeted region and TBI sequelae, as shown in Table 3, the analyzed studies invariably reported positive results. Animals were generally kept awake during DBS, unless the stimulation was applied simultaneously with cystometric assessments [172‐174]. Some studies utilized a current-controlled approach with amplitudes ranging from 20 to 200 µA [166‐169, 175] or at 80% of the individual resting motor threshold [170, 171], while others applied voltages between 1 and 4 V [172‐174, 176]. A stimulation frequency of 50 Hz seemed to be effective in the treatment of bladder dysfunction [172‐174], while lower frequencies were used to treat motor [171] and cognitive deficits [166‐168, 170], and higher frequencies of up to 200 Hz can be employed to increase arousal [175, 176]. Task-matched stimulation at 130 Hz for 5 s after each successful trial in a spatial learning test was also used to treat cognitive impairments after TBI [169]. In two studies, stimulation was applied directly before cognitive tests [166, 168], while, in the treatment of bladder dysfunction, stimulation was only triggered during cystometry when the measured bladder pressure exceeds a certain threshold [172‐174]. Three studies applied stimulation over 12 daylight hours over several consecutive days to improve spatial memory [167, 170, 171], and two others investigating the potential of DBS to increase arousal started their continuous stimulation protocols directly after TBI over the course of 2 h to 1 day [175, 176]. The targeted brain area and stimulation onset highly depend on the treatment application in question, since DBS can be used to stimulate relatively small brain regions—compared to other stimulation methods—without affecting the surrounding tissue. Long-lasting effects of DBS were only reported in [169], where researchers observed improved recovery of spatial memory 10 days after cessation of stimulation compared to untreated animals; meanwhile, other studies reported that they did not find persisting effects on hippocampal theta power after stimulation was terminated [166, 168].

VNS has been used in the preclinical studies listed in Table 4 to improve motor and cognitive impairments [177‐179] as well as disorders of consciousness [180, 181] after TBI, but also in the treatment of cerebral edema [182, 183] and to prevent cell death [184]. Animals were usually awake during VNS, except in two studies where researchers intentionally anesthetized animals to investigate the effect of VNS on disorders of consciousness [180, 181]. One study does not state clearly whether animals were anesthetized during the VNS or not [183]. Four studies applied stimuli at an amplitude of 0.5 mA and a frequency of 20 Hz [177, 178, 182, 184], while three other studies used currents between 0.8 and 1 mA with a frequency of 30 Hz [179‐181, 183], all of which chose to stimulate the left vagus nerve at the cervical level. Stimulation was often applied for 30 s in 30 min intervals over a period of up to 2 weeks, starting within 2 [177, 182] or 24 h after injury [178, 184], while two studies applied the stimulation only once, directly after induction of TBI [180, 181]. In one of the studies, stimulation was applied for 500 ms within 45 ms after each successful trial in a pull performance task, with the aim to improve motor function [179]. Only in one study stimulation was applied to the right vagus nerve at a frequency of 5 Hz with 5 ms pulses and an amplitude of 10 V, once for 20 min, in an effort to alleviate brain edema [183]. Most of the studies in this scope did not investigate any possible persisting effects, since VNS is mostly used as a continuous treatment after injury. The study conducted by Pruitt et al. measured persisting effects 1 week after the completion of VNS treatment; nevertheless, animals underwent further rehabilitation [179]. Two studies each observed that VNS attenuated the development of brain edema [182, 183], that it is effective for the treatment of cognitive [177, 178] or motor impairments [178, 179], had neuroprotective effects [183, 184], and promoted wakefulness after TBI [180, 181].

Experiments with TMS in preclinical models of TBI attracted interest rather recently with the oldest study dating back to 2015. All of the analyzed TMS studies in this scope employ rTMS protocols for effective treatment. Given that the early phases after TBI are associated with cortical hypoexcitability [185, 186], high frequency rTMS has been the major focus of interest in the studied publications. This is in line with the treatment window in these studies, which often starts relatively soon after TBI. On the other hand, low frequency rTMS induces inhibitory effects, rendering neurons less likely to fire [82], and is mostly utilized in epilepsy research [187]. It should be noted that post-TBI hyperexcitability is also observed, though after some time with an onset after approximately 2 months in preclinical models [188] and it is associated with trauma-induced epilepsy. Notwithstanding, preclinical experiments with low frequency rTMS for the prevention of TBI-induced epileptogenesis are currently quite underrated and further research is needed.

The inclusion of appropriate control groups in TMS studies deserves critical emphasis. Verdugo-Diaz et al., for instance, showed that movement restriction alone, which is necessary for stimulation in awake animals, significantly reduced post-traumatic bleeding and mortality, and improved neurobehavioral scores to the same extent observed in the rTMS group [157]. Similarly, combination of rTMS with environmental enrichment (EE) reportedly led to improvements in sensorimotor function lasting up to 6 weeks compared to the rTMS alone [154]. However, in this study rTMS was applied for only 1 week post-TBI, whilst EE lasted for 6 weeks. Unfortunately, both untreated TBI and TBI + EE controls were not included in the beamwalk tests, leaving the question unanswered whether rTMS itself had any long-term contributions to the observed improvement.

Large variabilities in the used stimulation frequencies (2–40 Hz), stimulation durations (3–20 min), treatment periods (a few days to 4 weeks), as well as heterogeneity in the used protocols for pulse trains, make a direct comparison between these studies difficult. Stimulation parameters were either taken from previous studies investigating modes of injury other than TBI [152, 155‐158], from clinical studies [153], or the choice of parameters was not mentioned [138, 154]. No two studies utilize comparable stimulation intensities, thus, a correlation of the stimulation parameters to different outcomes is hindered. Nevertheless, several studies with different TBI models, namely weight drop and controlled cortical impact (CCI), showed functional improvements upon rTMS starting 1 or 2 days after TBI, [138, 154‐157] with daily sessions usually administered for 1 week or longer. However, in a rat model of lateral fluid percussion injury (FPI), rTMS starting 4 days after induction of severe TBI did not show any improvements in motor behavioral outcome [152], whilst in a CCI model of pediatric TBI beneficial effects were reported after starting rTMS 9 days post-injury [153]. Similar improvements in neurological scores were also reported after moderate TBI using Feeney’s weight drop model, when rTMS was started 4 days post-injury [158]. Reported cellular and molecular biological readouts suggest that the observed functional improvements could be the result of neuroprotection, thus a critical time window for the treatment after TBI can be presumed. However, the existence of such a therapeutic window, and whether it is influenced by factors such as age, gender, and trauma severity, is unclear due to the limited number of published preclinical studies on this topic as well as the large variability in used parameters and treatment regimens.

Biological correlates of observed functional improvements could include mitigation of apoptotic signaling and cell death [152, 156], as well as reduced loss of mature neurons [155, 156] and astroglial activation [155] together with increases in cell proliferation and neurogenesis in the neurogenic niches such as the subventricular zone of lateral ventricles [156]. Moreover, upregulations in the expression levels of brain-derived neurotrophic factor (BDNF), tropomyosin receptor kinase B (TrkB, neurotrophin receptor), N-methyl-d-aspartate receptor 1 (NMDAR1, glutamate receptor) and phosphorylation of cyclic AMP response element binding protein (CREB; induced by neuronal activation) [158] support the presumption that restoration of cortical excitability early after TBI has a critical role not only in attenuation of delayed loss of cells that survived the primary impact, but also in the enhancement of regenerative responses. These results are of peculiar importance for a better understanding of underlying biological correlates of improvements that were detected in clinical applications, as most of these readouts are devoid of any possibility of direct assessment in the clinical practice. Whilst the positive results are per se encouraging—despite large variabilities in injury type, trauma severity and stimulation parameters—the translational value of preclinical studies is invariably dependent on their power in delineating correlative and causative relations between the applied stimulation parameters and observed biological readouts. Therefore, maturation of preclinical research on post-TBI rTMS from the current exploratory phase towards standardized procedures that allow for systematic comparisons is highly desirable.

Similar to TMS, tDCS has only been under investigation in preclinical studies as a possible treatment for TBI sequelae in recent years, starting in 2016. Several of the selected studies investigated the same TBI sequelae and how tDCS could be used to treat them in a preclinical setting. Two studies from the same research group assessed the effect of tDCS on microvascular cerebral blood flow (mCBF), starting in the chronic phase 3 weeks after trauma induction using either repeated or single stimulation sessions [161, 164]. They could observe a restoration of impaired cerebrovascular reactivity to hypercapnia, improved cerebral blood flow and tissue oxygenation, which is a key factor in brain metabolism associated with brain damage in the acute phase. A decrease in blood flow regulation together with decreased tissue oxygenation is suspected to cause damage in the early phase post-injury. Moreover, a chronic reduction of local brain perfusion in patients with TBI is known to cause persisting effects on brain function [189] and is thus suspected to play a crucial role in long-term outcome. An improvement in motor function and excitability could be observed in response to a single tDCS session directly after TBI induction [160] or in the subacute phase 24 h after the injury [163]. The beneficial effect of the stimulation was apparent from the day after the stimulation in both experiments and up to 12 days later, where the experiment ended. In two other studies, the improvement in motor function in response to repeated tDCS over 4 weeks in the chronic phase was examined. The first of these studies, where stimulation was started 3 days after trauma, observed a significant difference to the sham-stimulation group from day 8 to day 26 post-injury [162]. In the second study, two groups with different time points of stimulation onsets, 1 and 3 weeks after injury, were compared [161]. The group with later onset of stimulation made a better recovery than when stimulation started 1 week after TBI, showing that tDCS led to a significant decrease in neurologic impairment and an increase in motor function, memory and learning. This finding was in part supported by another study, where tDCS was started either 1 or 2 weeks post-injury and lasted for 5 consecutive days [159]. Here, the results of the Rotarod test in the 2-week group were slightly better than in the 1-week group. However, the observed improvement in spatial memory was comparable in both groups. Long-lasting or persisting effects after the end of stimulation were assessed in four studies. The first showed a beneficial effect on motor function and spatial learning directly after tDCS sessions ended, however, 2 weeks later the animals in the other groups had recovered to a similar level [159]. The second study could show a persisting effect of the stimulation 1 week after the end of the treatment [161]. The third study investigated structural brain damage in MRI 12 days after the trauma immediately after tDCS, but did not find any significant volumetric changes such as hydrocephalus or cortical thinning in either of the groups (sham, repetitive mTBI, and repetitive mTBI with tDCS). Immunohistochemistry did not show any evidence of neuronal degeneration in sham, TBI or stimulated group. Immunohistochemical study with glial fibrillary acidic protein (GFAP) stain showed a slight hypertrophy of cell bodies and a minimal extension of cell processes in both the TBI and anodal tDCS group compared to the sham group 12 days after the trauma [160]. Another study, where stimulation was applied for 7 days starting 6 weeks after TBI, found no persisting effects after stimulation was stopped [165].

These findings lead to the conclusion that tDCS decreases the time needed for recovery. From the evidence presented above, it is unclear if tDCS is able to induce persisting changes in neuronal tissue, although an improvement of motor function and tissue oxygenation could be observed over several weeks. The effect of anesthesia on the treatment outcome is not apparent in the selected studies and the choice of anesthetizing animals during stimulation is not directly correlated to the impairment under investigation. Most studies adopted their stimulation parameters from research papers that treated impairments caused by something other than TBI [159, 163, 165] and two gave no specific reasoning for their choice of stimulation parameters [160, 161] and later reused them in publications for further investigations [163, 164].

Follow-up studies could focus on investigating changes to the established stimulation protocols and how these changes affect treatment outcome, while using electrodes with standardized surface areas or adjusting the amplitude of applied currents to reach comparable current densities. The timing of stimulation onset seems to be an important factor for a better treatment outcome, however, there are no commonalities concerning the optimal time point for the start of stimulation post-injury among these studies. Cathodal stimulation is rarely used in tDCS studies, even though it was shown to be an effective treatment to decrease impulsivity and increase attention after TBI [165], and there may be additional applications for it. Most of the studies in this scope assessed the histologic changes after TBI and tDCS treatment, which could serve as a solid basis for further research into the therapeutic mechanisms of tDCS.

DBS first started to find use in preclinical studies about TBI treatment in 2013. The studies selected for this review used widely different stimulation protocols and time frames for each potential treatment application, which makes a comparison between them difficult. Almost half of preclinical DBS studies applied electrical stimuli continuously for 2 h [176], 1 day [175], or several weeks [167, 170, 171]. One research group initiated DBS whenever a signal measured via external urethral sphincter electromyography exceeded a certain threshold, in an attempt to enhance voiding efficiency [172‐174]. Another group started stimulation directly before an experimental task in order to improve cognitive outcome [166, 168], while in one study stimulation was triggered every time a rodent successfully found a hidden platform in a Morris water maze test, with the goal to reinforce learning [169]. Most of the time, animals received stimulation in multiple sessions over several days [166, 168, 169, 175] or weeks [167, 170, 171], with others only applying a single session before the animals were sacrificed for further analysis [172‐174, 176]. For the treatment of decreased arousal and disorders of consciousness, stimulation was usually initiated shortly after injury [167, 176], while treatment of bladder dysfunction started 1 week after induction of TBI [172‐174]. Therapy of cognitive deficits was shown to be effective in the acute [167], subacute [166, 168] and chronic phases of TBI [169, 170].

Two studies used higher frequency stimulation of 100 Hz or more in the thalamic region to increase excitability in animals suffering from decreased arousal [175] or disorders of consciousness [176]. Stimulation frequencies as low as 7.7–8 Hz were applied in the midbrain or medial septal nucleus to treat cognitive deficits [166‐168], while 30 Hz stimulation in the lateral cerebellar nucleus was used for a similar purpose [170, 171]. All three studies investigating DBS as a treatment for bladder dysfunction in this scope originate from the same research group and used identical stimulation parameters [172‐174]. Their triggered approach consists of 10 s of 50 Hz stimulation at amplitudes between 1 and 2.5 V. In their most recent study [174], they explored simulation of the pedunculopontine tegmental nucleus instead of the rostral pontine reticular nucleus to investigate its neural connectivity with bladder function, resulting in a similar outcome. Aronson et al. applied 130 Hz biphasic pulses in trains of 500 ms in the nucleus accumbens, whenever an animal succeeded a given task, leading to an improved spatial memory in TBI rats [169]. Only one group reported that they found no beneficial effects after stimulating the medial septal nucleus at a frequency of 100 Hz [168]. While one study did undocumented preliminary research to find optimal stimulation parameters [166], others adopted their parameters from previous studies on different topics [171, 172, 175, 176] or made the selection and optimization of the stimulation protocols part of their study [167, 168, 173, 174]. Jen et al. found an ideal stimulus length and frequency for effective stimulation for their purpose [172], only to continue with investigations regarding the optimal stimulation intensity in further studies [173, 174]. Only Aronson et al. do not describe how they chose the exact stimulation parameters they use, but mention that phasic stimulation in the nucleus accumbens might be able to promote neural plasticity [169].

Three studies found that DBS in various locations can be used to improve motor function after TBI [167, 171, 175] and three others observed an improvement in voiding efficiency [172‐174]. Two studies found that DBS improved spatial working memory [166, 167] and attenuated hippocampal theta activity [166, 168]. In one study, researchers observed a mediation of anti-apoptotic and anti-inflammatory effects after DBS [171], while another confirmed that it may promote wakefulness [176]. Most studies did not investigate any persisting effects of DBS. However, one study observed that the beneficial effects of their task-matched stimulation approach on spatial memory persisted 10 days after stimulation cessation [169], and several clinical studies have shown before that DBS leads to long-lasting positive changes in connectivity [190‐192]. Animals were usually awake during DBS, except in studies involving cystometric measurements where they were anesthetized [172‐174].

Researchers should continue building upon the insights gained in these studies about DBS as an effective preclinical treatment for TBI sequelae to find out more about the underlying mechanisms pertaining to precise electrical stimulation of specific brain areas. It would be desirable to find a consensus about the most effective stimulation parameters and time frames for a variety of impairments by comparing the effects of small parameter changes, as it was already shown in some studies in this scope. Experiments often lasted for less than 1 week, and animals were often sacrificed directly after an experiment or shortly after stimulation was terminated, having left no room for investigations into possible long-term improvements. Since DBS is used as a long-term treatment in clinical studies [114], preclinical studies should also address the effects of long-term stimulation. It remains to be seen if different impairments with related underlying neurologic causes may be treatable with similar stimulation protocols by stimulating in different brain regions.

Compared to other stimulation modalities, the efficacy of VNS in the preclinical treatment of TBI sequelae has been investigated for a much longer time since 2005. Almost all VNS studies in this scope stimulated the left vagus nerve at the cervical level, except for one that targeted the right vagus nerve [183]. This consensus may stem from the fact that the right vagus nerve has more projections to the sinoatrial node of the cardiac atria and stimulation could therefore have an undesirable effect on the cardiac rhythm [130].

In the analyzed publications, most research was focused on treatment with multiple repeated stimulation sessions [177‐179, 182, 184], while a few publications report the effects of single session VNS treatment [180, 181, 183]. The onset of the treatment in studies using repeated stimulation varied between 2 h [177, 182], 24 h [178, 184] and up to 9 days [179] after sustaining TBI. These time points correspond to different phases of post-injury pathology: early acute phase, subacute phase and chronic phase. In a clinical setup, therapy can be implemented at any point after TBI; however, early interventions are known to lead to better functional and psychological outcomes in patients [193‐196]. Additionally, the long-term study of Pruitt et al. delivered stimuli within 45 ms after each successful pull trial [179], which should lead to strengthening of synaptic connections according to the STDP model of plasticity. In the studied publications, an early onset of the VNS treatment led to a faster recovery of motor skills, which is usually observed around day 2 [177, 182], as compared to a subacute onset from day 4 on [178]. Conversely, starting stimulation 24 h after TBI encouraged a faster improvement of cognitive functions; 13 days for early-onset [177] compared to 11 days for the later-onset study [178]. The study implementing VNS in the chronic phase also showed a positive effect of VNS on motor recovery [179]. However, it was sustained for 5 weeks and accompanied by physical training of the animals. Persisting effects of VNS were described for 1 week after cessation of the treatment. Multiple stimulation sessions also proved to have a neuroprotective effect on GABAergic neurons [184] and limit edema formation in the ipsilateral cortex [182]. In summary, repeated VNS aids in functional recovery after TBI and to some extent helps in constraining secondary damage.

Single stimulation after TBI led to a wake-promoting effect in free-fall injury animal models [180, 181] and the promotion of anti-inflammatory cytokine modulation with lower edema formation in a blast injury model [183]. This might indicate that an isolated VNS session could be advantageous in early post-injury stages and might lead to diminishing secondary injury. Nevertheless, clinical application of single VNS treatment would be plausible in the form of non-invasive stimulation, rather than during surgery. Transcutaneous VNS has already been proven feasible and was well tolerated in humans with severe TBI [197]. Pre-clinical studies employing this kind of VNS treatment for TBI are not available at this moment, but have been described for models of inflammation [198, 199], ischemia [200] and seizures [201].

Unlike in the case of TMS, the VNS studies in this scope use comparable stimulation protocols. Four publications coming from one research group report using the same stimulation parameters, which authors described that they were adapted from a previous study [177, 178, 182, 184]. This leads to a better reproducibility of the experiments and facilitates comparability of the results. Other studies mention implementing the same stimulation protocols as previous publications where the respective impairments had a different underlying cause than TBI [179‐181], while one study does not mention how stimulation parameters were selected [183].

Since VNS is an established method and was FDA-approved for drug-resistant epilepsy and depression [202], there is an abundance of commercially available devices for human patients. However, similar devices for rats are currently not accessible and only some of the publications [177, 178, 182, 183] describe the electrodes they were implanting. Post-experimental re-testing of the electrodes is reported in only three of them [177, 178, 184]. None of the studies mentions pre-surgical evaluation of the devices, which might be crucial to ensure proper functionality. Similarly, observed side effects were also not reported in any of the analyzed publications, which could lead to insights into safety of VNS application in patients.

Since VNS is usually used as a long-term treatment in awake patients, the effect of anesthesia on the stimulation outcome is not investigated, unless it is specifically used as a treatment for disorders of consciousness [180, 181]. However, these studies report usage of a chloral hydrate, a drug considered not suitable for anesthesia of laboratory animals [203], and mention inducing anesthesia three times during 1 day in some of the experimental groups, which is a considerable burden for animals. Therefore, these results should be interpreted with caution.

Compared to the other stimulation methods presented above, there is more consensus between different VNS studies. This method proved to be advantageous for therapy of different conditions associated with TBI, regardless of the temporal window of its implementation and the amount of stimulation sessions. Further studies aimed at different modalities of VNS, e.g. transcutaneous VNS, and combination with other therapeutic agents, such as physiotherapy and pharmacotherapy, as well as life-long studies might lead to additional insights into better applications of VNS in humans.

All of the methods discussed here can be used to treat motor and cognitive dysfunctions and lead to significant improvements in TBI animal models [138, 153‐155, 159, 160, 162, 163, 166‐171, 177‐179]. Only one study found that their TMS protocol did not induce any beneficial effects regarding motor improvements [152], which was likely due to the relatively short stimulation duration they used compared to other studies. At the same time, there are a variety of other TBI sequelae that benefit from treatment with different electrical stimulation modalities. Neuroprotective effects can be induced with TMS, DBS and VNS to prevent further cell-death after injury [153, 155, 171, 183, 184]. Both TMS and tDCS are able to modulate cortical excitability leading to plasticity and increased brain activity [153, 154, 160, 165]. Suppression of cortical excitability can be achieved with TMS and tDCS as well, leading to a decrease in hyperactivity and impulsivity in animals [153, 165]. After stimulation with TMS and DBS, researchers discovered beneficial changes in histological assessment [157, 158, 168‐171], while some tDCS and VNS studies show positive effects on protein expressions after treatment [159, 160, 162, 180, 183]. Finally, the studies in the scope of this review show that tDCS, DBS and VNS may effectively be used to promote wakefulness and treat disorders of consciousness caused by TBI [160, 175, 176, 180, 181].

While these stimulation methods have many treatment opportunities in common, each of them have possible applications that have not yet been observed with the other modalities in TBI animal models, giving them a status as some sort of “specialization”. TMS has been used to improve brain metabolism and potentially induce cell proliferation and neurogenesis [156], while tDCS studies showed that it can be used to increase microvascular flow and tissue oxygenation [161, 164]. This likely stems from the fact that these two methods both activate large parts of the cortex, therefore having a higher impact on the metabolism and oxygenation of the brain. Exclusively, DBS studies explored the application of electrical stimulation to improve voiding efficiency in animals with bladder dysfunction [172‐174], since DBS can be used to specifically target diseases whose etiology is connected to single brain regions. Only VNS, which is known to decrease the disruption of the blood–brain barrier [120], has been used in preclinical studies to attenuate the development of cerebral edema after TBI [182, 183].

There are several aspects of pre-clinical studies that should be taken into consideration while analyzing their translatability into a clinical environment. Among them worth mentioning are: the relevance of the animal model, appropriate treatment, the temporal window, and side effects.

All of the analyzed studies were performed with well-established mammalian model species: rat, mouse and rabbit. The most frequently used model species was rat (28/34), with Sprague–Dawley as a leading strain (19/28), followed by Long Evans rats (7/28) and a single instance of Wistar rats used. A minor portion of analyzed studies was performed on mice (5/34) and only one publication reports experiments on New Zealand rabbits. The dominance of the rat model stems from a relatively big size of the brain in these animals, as compared to mice. This translates to convenience during surgery, especially when small electrodes are implanted, but is also important for a better spatial resolution when targeting specific brain regions [204], for instance with TMS. Common use of Sprague–Dawley rats ascertains comparability of the results within and between neurostimulation methods. However, Sprague–Dawley rats were reported to reach a faster motor skill recovery as compared to Long Evans rats [205]. Therefore, caution is recommended when comparing the two strains with each other. Moreover, all of those species are lissencephalic and display different geometry, craniospinal angle and grey-to-white matter ratio than humans [206], which is a further limitation of the translatability of results to human patients. Only one publication [153] used juvenile rats as a model for a TMS study. Since TBI is the disease with one of the highest incidences in children and youth below 19 years old [207], it is of utmost importance to further encourage studies employing neurostimulation methods as a post-traumatic therapy in young animals, with special focus on non-invasive methods.

Sex-dependent differences in the outcomes of TBI pre-clinical studies have been widely reported and reviewed in multiple studies [208‐210]. In general, animal studies report better outcomes in females than in males, which might stem from the neuroprotective effects of estrogen and progesterone [208, 209]. The desire to determine treatment efficacy independent of hormonal status leads to the selective inclusion of males in pre-clinical studies, unless the study is specifically designed to address the sex difference itself [208, 209]. Likewise, only a small proportion of publications analyzed here reports use of female animals (2/34) [172, 179] or both sexes (2/34) [176, 180], which restrains the translatability of the results into human patients. Inclusion of female animals in experimental post-traumatic neurostimulation research is recommended for a better representation of the clinical situation.

Methods of inducing TBI varied: 15/34 studies used a weight-drop method, 10/34 fluid percussion injury, 8/34 controlled cortical impact and 1/34 performed blast injury. Except for blast injury, which is not fully consistent, these models are highly standardized and cover different types of injury, from focal to diffuse and mixed [206], corresponding to lesion diversity in patients who have survived head injury. The reported severity of the TBI model also varies: mild TBI was induced in 3/34 of studies, mild-moderate TBI in 2/34, moderate TBI in 13/34 and severe in 5/34. This does not fully mirror the clinical situation in humans, where approximately 80% of TBI is categorized as mild [211]. Nevertheless, moderate and severe TBI constitute approximately 50% of hospitalizations [212] and lead to higher mortality [213, 214]. Regrettably, a significant proportion of studies (11/34) does not report the severity level of the injury, substantially limiting their translatability. It is also worth noting that, due to anatomical and coil size differences, TMS may be able to stimulate deeper brain regions in small rodents that could otherwise not be effectively targeted in human patients [25].

Appropriate treatment requires a suitable method and stimulation protocol for the disability under investigation in the respective study. This is especially an issue for the clinical applicability of the TMS, tDCS and VNS studies in the scope of this review, since they use widely different stimulation parameters and time frames, even for the treatment of similar disabilities. In case of VNS, stimulation protocols were comparable; however, studies performing acute VNS intraoperatively might not be as clinically relevant.

The temporal window of applied stimulation methods varies highly. Early onset of the stimulation protocol was reported in 9/34 of publications analyzed, subacute in 13/34 and chronic in 13/34 of studies. Interestingly, individual methods seem to be applied at specific time points: TMS almost exclusively in the subacute stage, tDCS mostly in the subacute and chronic stages, and VNS in the acute and subacute stages, while only DBS finds application in all stages after TBI. This distribution of the time points may correspond well to the clinical situation, in which onset and duration of the therapy are highly variable [215‐217].

Finally, possible adverse effects of the treatment are an important factor as well. The presence of side effects during pre-clinical studies might indicate plausible future problems in the clinical setting and should not be underestimated. Review articles on side effects caused by therapeutic application of TMS, tDCS and VNS in a clinical setting report only mild side effects [218‐220], while adverse effects of DBS require more investigation in closer cooperation of scientists and clinicians [221] and are prone to bias [222]. Very few of the studies in this scope investigated possible side effects of any of these four stimulation methods and not a single one reported that they found any negative implications, which similarly hints to a possible bias and would be an important aspect in further research on this topic.