Transcranial Doppler Ultrasound and Concussion–Supplemental Symptoms with Physiology: A Systematic Review

Study selection and characteristics

A total of 141 articles were identified by the PubMed database, with 125 of these being excluded during the title and abstract screening (Fig. 1). Of these, exclusion reasons consisted of the articles not being original research articles (n = 14), studies not being conducted in humans with SRC (n = 71), and studies using neuroimaging techniques other than TCD (n = 40) (Fig. 1). The 16 remaining articles were screened during the full-text stage with two articles being excluded as they were not original research articles (Fig. 1). A total of 14 articles were therefore included in the qualitative synthesis (Fig. 1).

OPEN IN VIEWER

FIG. 1.

A modified Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) diagram detailing the number of studies screened across each phase of the systematic review.

All 14 included studies were conducted in North America, 8 were completed in the United States (57%), and the remaining 6 were conducted in Canada (43%). Studies were published between the years of 2011 and 2021, with an average sex breakdown of 35% females and 65% males. Seven studies stated the competition level, of which three were completed in both elite junior hockey and American football, one in various college sports, and three in various recreational sports including soccer, rugby, and hockey. Five studies were completed in youth athletes (ages 5–12), one in adolescent athletes (ages 13–17), and eight in a mix of both young adults (≥ 18 years of age) and adolescents. The most common guideline used for concussion diagnosis from the healthcare providers was the current Concussion in Sport Group consensus statement at the time of study enrollment. ^1,29 Across the studies, one included an assessment of the neurovascular coupling (NVC) response, seven quantified cerebrovascular reactivity (CVR), five assessed dynamic cerebral autoregulation (dCA), and one study examined both CVR and dCA.

Risk of bias assessments

The SIGN rating revealed that seven of the articles were deemed to have a high risk of bias, eight were deemed acceptable, with 0 high-quality study designs (Table 1). The average MINORS score for articles without a control group was 11/16, and was 18/24 for those with a control group. The ratings for each article are shown in Table 1. A common source of bias in many studies included the lack of control for known confounding influence, where only six (40%) ^{30 –35} of the studies specifically stated that participants abstained from caffeine, exercise, alcohol, and/or smoking for 12–24 h prior to study commencement. These factors are known to impact cerebrovascular regulation ^{36 –39} and therefore are factors that can minimize the internal study of a study design. Only one study included an a priori sample size calculation. ³² Although numerous studies included female participants, none of these controlled for phase of the menstrual cycle and/or hormonal contraceptives, which may influence cerebrovascular function. ^40,41 Additionally, the studies that quantified adolescent athletes did not report on pubertal status and/or development age. Finally, the control groups used across studies were very heterogeneous, ranging from asymptomatic, unmatched controls, to age- and sex-matched controls, or physically active controls, and some studies compared SRC athletes to their own pre-injury baseline values.

Neurovascular coupling

Only one article included in the present review used TCD to assess the NVC response (Table 2). This study utilized a longitudinal case series study design, in which participants were assessed at three time points (i.e., 72 h, 2 weeks, and 1 month) and compared with their own baseline values. ³³ The study insonated the left posterior cerebral arteries (PCA) and right middle cerebral arteries (MCA) of 18 participants. NVC was observed in response to a reading stimulus on a computer screen, alternating between 20 sec of eyes closed and 40 sec of the visual tasks with eyes open. ³³ Wright and colleagues ³³ found no differences regarding the absolute values in the PCA and the MCA pre- and post-SRC, or in the cross-sectional aim comparing those with and those without a history of concussion. However, the PCA velocity maximal amplitude of response was not only delayed but also required increased CBv levels for the same task in the acute post-SRC period compared with per-injury baselines, which remained affected at the 14-day time point. Further, medical return to play was inversely associated with the magnitude increase of the NVC response at the 72-h time point. No correlations between NVC response and Sport Concussion Assessment Tool 3 (SCAT3) symptom score, Standardized Assessment of Concussion (SAC) score, or Balance Error Scoring System (BESS) score were observed. ³³

CVR

Eight included studies investigated cerebrovascular function in a SRC population using a vasoactive stimulus (Table 3). Studies investigated concussed populations of varying sizes from n = 10⁴² to n = 70. ⁴³ Seven studies employed a comparison group that was not concussed ^{31,42 –46} or was asymptomatic/recovered from a recent concussion, ⁴⁷ whereas one study did not use a comparison group at all. ³⁰ Studies ranged from cross-sectional, single time point assessments ^42,46 to up to 10 assessments on a single participant. ⁴³ All studies used TCD to assess MCA velocity, with unilateral assessment on the right side (n = 4), ^30,42,44,47 unilateral on either side (n = 1), ⁴⁶ bilateral (n = 1), ³¹ or unspecified (n = 2). ^43,45 All studies used a hypercapnic stimulus, with three also completing a hypocapnic protocol. ^30,31,42

Hypercapnia was achieved by re-breathing room air in a 5 L bag, ^44,46,47 re-breathing a mixture of 5% CO₂, 50% O₂, and 45% N₂ from a 10 L bag, ⁴⁵ completing repeated breath holds, ^30,42,43 or through the inspiration of an unspecified hypercapnic gas mixture. ³¹ . One study (re-breathing from a room air-filled bag) reported an association between increased vasoreactivity and more severe headache; however, the outcome metrics were not significantly different between concussed participants and controls. ⁴⁷ Another study (re-breathing from a hypercapnic-filled bag) reported lower CBv, respiration, and lower exercise tolerance in individuals experiencing persisting post-concussion symptoms (PPCS) compared with controls. ⁴⁵ Further, Len and coworkers. (20 sec breath hold) noted a difference in the CVR response acutely following SRC (2 days), ⁴² which was corroborated by a future study of the same group, (breath hold) but only at the time point 2 days post-injury, and not at 1, 4, or 8 days post-SRC. ³⁰ This was contrary to the study by Thibeault and colleagues (25 sec breath hold), which reported no difference in CVR during the first 48 h post-SRC. ⁴³ Purkayastha and coworkers (hypercapnic inspiration) discovered CVR deficits at the symptomatic stage, which remained once athletes returned to play. ³¹ Further, Aaron and coauthors (re-breathing in a 5 L bag) reported higher vasoreactivity in a concussed population; however, this appeared to be influenced by three outliers in the concussion group. ⁴⁴ Lastly, Howell and coworkers (re-breathing in a 5 L bag) established that resting CVR performance was associated with cerebrovascular response to exercise-induced changes in the end-tidal partial pressure value of carbon dioxide (P_ETCO₂) within the concussed group. ⁴⁶

In the studies that quantified the hypocapnic CVR response, hyperventilation was used to achieve ∼22 mm Hg, ³¹ or a rate of 36 breaths/min without a target P_ETCO₂. ^30,42 Two studies found no significant changes in CVR to hyperventilation between baseline and all three follow-up time points, ^30,42 whereas one did not report separate CVR calculations between hypocapnic and hypercapnic trials. ³¹

Cerebral autoregulation and orthostatic intolerance

Six studies included in the systematic review assessed dCA with sample sizes varying from 6⁴⁸ to 37⁴⁹ in the SRC group with up to 179 individuals, ³⁴ including baseline measures (Table 4). Of these studies, assessment time points varied from single assessments ^32,48 up to 6 different testing days. ⁴⁹ All studies insonated the MCA, with three completing right unilateral assessments ^32,49 and two completing unilateral assessments on either side, ^34,35 as well as one bilateral assessment. ⁴⁸ The protocols used to quantify CA included the use of sit-to-stand maneuvers, ⁴⁹ a head-up tilt test, ⁴⁸ squat-stand maneuvers, ^34,35 and lower body negative pressure (LBNP). ³² Further, one study used an LBNP -40 Torr and a lower body positive pressure (LBPP) of +20 Torr to passively create a prolonged orthostatic challenge. The outcome metrics from these studies included rate of regulation, ⁴⁹ autoregulatory index, ⁴⁸ transfer function analysis, ^34,35 projection pursuit regression, ⁴⁴ and the absolute and relative changes in CBv. ³² Despite the heterogeneous methods and outcome measure, all studies noted CA impairments acutely post-SRC. ^{32,34,35,48,49} Two studies completed subsequent evaluation at an asymptotic time point, noting physiologically based CA impairments persisting even when participants were clinically cleared for return to play. ^34,35 Further, the study by Worley and colleagues who assessed orthostatic intolerance noted that concussed athletes had a lower total LBNP time and had a greater change in MCA during the LBNP protocol. ³² Finally, the study that performed bilateral MCA assessment observed that five out of six participants experienced impaired autoregulation, three out of six experienced lower MCAv, and two out of six had higher resting MCAv. In those who sustained a unilateral mild traumatic brain injury, three out of six had impairment observed in the MCA ipsilateral to the location of the brain injury. ⁴⁸

Aim 1: Review current literature examining the acute and subacute effects of SRC on intracranial hemodynamics

Concussion is known as a “snowflake” injury because of the heterogeneous symptom and clinical presentations, produced in part by the variety of injury mechanisms and the diversity of affected cerebral regions. ¹ Nevertheless, the included studies within this review consistently noted deficits to the intracranial artery responses (primarily in the MCA) to metabolic (i.e., NVC), chemical (i.e., CVR), and systemic (i.e., dCA) influences, despite the fact the underlying physiology of these three regulatory mechanisms is quite distinct. ¹⁴ When an individual engages in a metabolic task that challenges a certain cortical region, this will result in an excitatory response through an efflux of glutamate. ⁵⁰ In the neuron, this will trigger the production of neuronal nitric oxide synthase (NOS). This is a precursor to nitric oxide and will result in a relaxation of the endothelial smooth muscle, subsequently causing localized vasodilation near the activated region of the brain. ^51,52 Additionally, in the surrounding astrocytes, arachidonic acid will be produced, leading to an increased concentration of epoxyeicosatrienoic acid and prostaglandin E2, and a dilation of the cerebral vessels. ^{51 –53} For a visual representation of the NVC response, readers are directed to Figure 3 in the article by Beishon and coworkers, which demonstrates the feedforward and feedback loops of NVC. ⁵⁴ According to the neurometabolic cascade of concussion, glutamate increases in the brain within the hours following injury, which will fall to ∼60% of baseline levels until physiological recovery. ^9,10 Given the energy mismatch, a state of hyperconnectivity is thought to occur where a greater metabolic demand is required to complete the same task compared with the pre-injury state. ⁵⁵ This phenomenon has been demonstrated through electroencephalography ⁵⁶ and is further supported by the findings by Wright and coworkers, ³³ who noted a greater total NVC response to the same visual stimulus following SRC (Table 2). The inflated CBv response for a given task persisted, despite clinical symptom resolution and return to baseline of SCAT measures. This will be further discussed in the next section. Ultimately, although more work is warranted to determine the precise underpinnings of the overreactive NVC seen following SRC, this is likely the result of the inefficient signalling of glutamate on both the neuron and astrocytic processes to control constriction/dilation. ^51,52

In contrast to the NVC response resulting in an excitatory/inhibitory response mediated by the neurons/astrocytes, both the CVR and dCA responses result from indirect or direct stimulus acting on the endothelial smooth muscles. ^57,58 During a hypercapnic CVR protocol, vascular accumulation of carbon dioxide will diffuse into the interstitial space. This will lower the pH surrounding the cerebral vessels as a result of carbonic acid dissociating into hydrogen ions (H+) and bicarbonate. ⁵⁸ The increased acidity activates nitric oxide production (via endothelial NOS production), resulting in relaxation and dilation of the endothelial smooth muscle. Conversely, hypocapnia reduces the amount of carbon dioxide, increases the pH, and thus will lead to an excitation and contraction of the cerebrovasculature. ⁵⁸ Figure 1 in the article by Liu and colleagues demonstrates the CVR response. ⁵⁸ With an increase in blood pressure, more tension (i.e., sheer stress) will be placed on the endothelial lining of the cerebrovasculature. ⁵⁷ This will result in a calcium efflux, with endothelial NOS being activated to produce nitric oxide and a subsequent dilation of the vascular wall. ⁵⁷ Under states of reduced pressure, the same relationship exists in reverse, where lower shear stress leads to vasoconstriction of the cerebrovasculature. ⁵⁷ Figure 1 in the article by Stauss and coworkers highlights the physiology of acute changes in blood pressure. ⁵⁹

Given that both CVR and dCA work through the endothelial NOS (eNOS) pathways, ^57,58 it is not surprising that similar deficits were found to both of these responses following SRC. Interestingly, the methodological approaches that used greater carbon dioxide or blood pressure perturbations noted greater and longer-lasting autoregulatory impairments (Tables 3 and 4). For example, the CVR studies employing breath-hold techniques noted mixed findings in the acute phase, with some reporting a difference, ⁴² others finding no difference, ⁴³ and one finding an isolated time point difference bookended by non-different results at time points either sides during the acute phase ³⁰ (Table 3). A prior systematic review by Gardner and colleagues ⁶⁰ referenced two studies that assessed CVR via alternating breath hold, identifying short-term impairments following SRC with no long-term effects after 5 days. However, some investigations in the present review with a more intense stimulus (e.g., re-breathing, inspired carbon dioxide) reported greater deficits following SRC (Table 3). One source reported no significant CVR deficit in a PPCS population ⁴⁷ despite another study linking CVR deficits with PPCS. ⁴⁵ Another study found persistent deficiencies in those whose symptoms had resolved and who had returned to play ³¹ (Table 3). Differences in CVR reporting among studies likely stems from the different CVR task used to quantify the hypercapnia response, as breath holding and inspired CO₂ challenges produce heterogeneous between-participant results (Table 3). This is attributable to these tasks eliciting differing levels of carbon dioxide accumulation, where all individuals do not have a homogeneous chemoreceptor sensitivity to a given stimuli. ⁶¹ However, the use of a more intense hypercapnic stimulus appears to detect asymptomatic cerebrovascular impairment that went unrecognized in the studies included by Gardner and colleagues in a previous review. ⁶⁰ Further, the included CVR studies noted deficits during hypercapnia challenges, whereas no differences were noted during hyperventilation protocols (hypocapnia) (Table 3). This finding is consistent with research noting that the acute impacts of exercise are isolated to hypercapnia. ³⁸

In contrast to CVR, all dCA studies reported impairments post-SRC (Table 4), despite heterogenous methodological and analysis approaches. Interestingly, the one study that noted persisting deficits despite symptom resolution employed squat-stand maneuvers ³⁴ (Table 4). These are known to result in large robust ∼30–50 mm Hg blood pressure oscillations compared with the other used techniques (e.g., deep breathing, sit-to-stand), which only elicit small to moderate oscillations (∼5–15 mm Hg). ^19,62 This corroborates the CVR findings noting that a greater methodological stimulus leads to potentially augmented discriminatory ability following SRC, which will be further discussed.

Aim 2: Consolidate the current application of TCD for assessing SRC recovery

A few of the investigations included in the current review also quantified cerebrovascular function after symptom resolution and return to play: NVC (n = 1), ³³ CVR (n = 5), ^{31,43 –45,47} and dCA (n = 2). ^34,35 Deficits to all three regulatory processes were found to persist, even with some studies noting return to baseline scoring for SAC and BESS scores at the 14-day time point. ³³ Return of physiological-baseline regulatory function was found for the NVC and dCA investigations at the 28-day assessment. ^33,34 Further, both studies completed analysis comparing athletes with no reported history of SRC and those with three or more past SRCs. No differences in baseline or post-injury cerebrovascular regulation were found between groups, suggesting that although physiological deficits may persist post clinical symptom resolution, pre-injury regulatory function of physiological parameters are eventually achieved. ^33,34 Nevertheless, the emerging literature highlights physiological deficits that remain following SRC despite symptom resolution and clinical recovery. ⁶³ Therefore, is it likely that the various methods of neuroimaging modalities summarized in this review quantify a different construct not captured by traditional clinical symptom reporting. However, this may be the result of the fact that neuroimaging studies using basic group statistical analysis with an alpha of 0.05 to identify the recovery of physiological processes may not be the most sensitive approach, as individual variation may wash out subtle changes. ^{64 –67}

As cerebrovascular recovery occurs, CBv differences between concussed populations and healthy controls will trend toward convergence with one another. Further, CBv is influenced by a multitude of factors such as caffeine, alcohol, and nicotine consumption, resting sympathetic tone, recent exposure to exercise. ^{36 –39,68} These factors have the potential to alter cerebral blood flow and intracranial arterial diameter, the latter of which is left unassessed by TCD. Therefore, TCD assessments have the potential to be “noisy,” and differences must be quite pronounced to be produce significant findings. Because of this physiological noise in TCD assessment, inadequate control of confounding factors may result in studies reporting cerebrovascular recovery prematurely. ^15,16 Therefore, more sensitive assessment strategies, employing multimodal assessments, diligent controls of sympathetic innervating factors, and more elegant statistical approaches must be employed before TCD can be considered as a clinical application to assess SRC. Further, there is evidence that there are long-term impairments in some cases post-SRC. ⁶⁹ For example, some investigations have highlighted that baseline concussion symptom reporting via the SCAT symptom evaluation is greater in those with a history of concussion than in those without. ⁶⁹ These symptoms commonly include headache, pressure in the head, fatigue, which may be underpinned by an autoregulatory deficit, but have not been detected because of “hamstrung” cerebrovascular assessments.

Symptom reports of feeling slowed down, fatigued, and/or drowsy upon medical clearance may in part be attributable to a state of hyperconnectivity in which greater neuronal activation is required to complete the same task. ⁵⁵ This would result in a greater glucose demand and subsequent production of metabolic byproducts such as hydrogen ions, carbon dioxide, and adenosine. ⁵¹ This combination has shown to produce fatiguing effects in healthy participants during sustained attention tasks. ⁷⁰ Symptoms such as headaches and pressure in the head may be related to dCA deficits, ^12,13 as this mechanism is challenged during daily tasks that elicit an increase in blood pressure such as walking up a flight of stairs, bending over, defecating. If there is an impairment regarding the proper signalling from an elevation in sheer stress to the release of nitric oxide and vasodilation of the endothelial lining, this could potentially contribute to a headache. ⁷¹ For example, an insufficient filtering of blood pressure increases may result in an overall increase in cerebral blood flow at any given time, which according to the Monro–Kellie Doctrine, ⁷² would result in an elevated intracranial pressure. ⁷³ Although no pain receptors are present in the brain parenchyma, the augmented pressure would be detected by the cranial meninges, ⁷⁴ resulting in a subjective presentation of headache/pressure in the head. Nevertheless, as stated, the studies included in this review noted full recovery of the autoregulatory processes when quantified via group averages. Further, assessment techniques such as TCD do not assess diameter, ⁷⁵ and as such, they are unable to directly assess the perturbation of vascular tone caused by SRC in response to a dCA, NVC, or CVR stimulus. Therefore, longitudinal studies with elegant assessment strategies, which would be more sensitive to subtle individual differences, are warranted to better understand if associations exist between specific symptoms and autoregulatory impairments.

Recommendations from Aims 1 and 2

Deficits to NVC, CVR, and dCA were present, but heterogenous, in the acute and subacute phase following SRC, with several studies noting persisting deficits despite symptom resolution and return to play, but others failing to observe a deficit. These investigations involving prolonged impairments generally used larger and more robust stimuli to further challenge the autoregulatory processes, which may have increased the sensitivity to detect subtle changes compared with baseline/controls groups. Although more research is warranted, these future investigations should consider using autoregulatory challenges that result in large physiological responses for each respective autoregulatory domain (i.e., NVC, CVR, and dCA). This will help unveil the specific time-course recovery of SRC on these deficits. It also appears that hypercapnic CVR stimuli are more sensitive at detecting SRC alterations than hypocapnic stimuli; therefore, a greater emphasis should be placed on the former. ³⁸ Finally, using regression/correlation analyses may increase sensitivity to determining autoregulatory impairments that would not be present during basic group comparisons with a binary p value. ^{64 –67}

Clinical relevance

The clinical utility of TCD following SRC lies in detecting impairments acutely following SRC and detecting impairments that persist following symptom resolution that would otherwise remain undetected. Because reliance on symptom resolution to determine return-to-play decisions risks exposing athletes to re-injury, ⁷⁶ the utility of detecting asymptomatic impairment is of utmost importance in this population. Present clinical use of TCD for use in traumatic brain injury, embolism, hemorrhage, vasospasm, sepsis, and pre-eclampsia is described in literature. ^{77 –79} However, compared with in these populations, the use of TCD in clinical settings for concussion, where several concerns must be addressed/considered before widespread implementation, is in its infancy.

First, Bhuiyan and coworkers ⁸⁰ demonstrated that the ability to obtain reliable TCD measures is heavily influenced by training, in that medical professionals without formal training do not perform reliable TCD assessments. Nevertheless, these authors demonstrated training programs can be used to remedy this issue. ⁸⁰ To maintain sufficient reliability of the cerebrovascular assessment, specifically trained ultrasonographers would ideally have to perform the assessments, as slight differences in insonation angle can produce vast differences in CBv. ⁸¹ In summary, improper assessments leading to differences in insonation angle are likely to lead to a misclassification bias of CBv and disguise any changes that occur following SRC.

Second, confounding effects of caffeine and alcohol consumption, circadian rhythm, blood pressure, ventilation (via end tidal carbon dioxide), and sympathetic tone are at risk of artificially attenuating changes in cerebrovascular function following SRC. Generally, these factors are easily controlled in laboratory settings; however, this becomes a greater issue in clinical settings. For example, utilizing TCD on a sideline setting at a sports field poses the potential for confounding influences of acute exercise from the sporting environment, acute caffeine/supplement consumption, and differing times of day. Reliable controlling of these factors is difficult to achieve in a clinical setting; therefore, further investigation as to the effect of each factor on cerebrovascular impairment following SRC is fundamental before clinical interpretation is possible.

Lastly, differences in age and sex will affect healthy cerebrovascular function, ^28,82,83 therefore making cerebrovascular dysfunction difficult to quantify without appropriate normative ranges. Future research establishing these ranges would be clinically relevant to many neurophysiological disorders. The evidence discussed in this review compared cerebrovascular impairment with either a pre-injury or post-injury “normal” baseline assessments of age- and sex-matched healthy controls. In the absence of normative data, the present research provides support for the diagnostic utility of TCD against a baseline reading. In summary, because of the difficulties in achieving a baseline cerebrovascular assessment outside of sport participation, the present evidence informs future directions for the development of TCD in clinical settings in sport.

Limitations and future directions

The major drawback of the current review is that the search was conducted exclusively in the PubMed database. This database, nonetheless, indexes all physiological, clinical, and biomedical journals that publish content relating to concussion and TCD. Additionally, to ensure that all potential articles satisfying inclusion criteria were identified, authors scanned the reference list of the included articles to determine if any citations were missed. However, no additional citations were unveiled.

Although TCD is an underutilized assessment technique in SRC in general, there was a distinct lack of studies providing secondary or exploratory analyses into potential sex differences. Therefore, ensuring adequate representation of females with sufficient power to perform sex-based comparisons is paramount. Additionally, only one investigation ³² completed an a priori power analysis, which serves to augment the internal validity of a given study. With respect to the three regulatory processes assessed, it appeared that using more robust stimuli unearthed greater differences between the SRC and control groups. Therefore, it is recommended that careful consideration be employed when developing the methodological design for a given study.

Previous work by Fierstra and colleagues ⁶¹ compared the benefits and limitations of different hypercapnic assessments to quantify the CVR response. Although breath-hold tasks require minimal equipment and are simple to implement, the use of these to assess cerebrovascular function has not been recommended. This task requires individuals to hold their breath for ∼20 sec. However, the transit time from blood in the pulmonary capillaries at the onset of this maneuver to reach the brain capillarization occurs between 10 and 20 sec. Therefore, although one may see a slight increase in hemodynamics during a breath hold of this duration, it is likely insufficient to produce a robust and consistent elevation. Further, this technique is incapable of producing a plateau at a desired P_ETCO₂ value. Another commonly used method is fixed inspiration with 5% or 7% carbon dioxide for 2–3 min. ⁸⁴ Although this method abolishes the mentioned transit time issue, a problem with this method is the inter-individual difference in central chemoreceptor sensitivity to carbon dioxide. Therefore, a 7% carbon dioxide stimulus may have a minimal impact on one individual, while causing a maximal CVR response in another. Therefore, this method may be applicable for within-group designs but faces limitations with between-group comparisons. To counteract the abovementioned issues, the use of end-tidal clamping has been suggested as the “gold standard” for CVR assessments. This technique has been conducted with different devices (e.g., RespirAct, AirForce) that seek to clamp the P_ETCO₂ at a specific target, independent of the individual's breathing frequency and tidal volume. ⁸⁵ These techniques are attached to gas cylinders of oxygen, carbon dioxide, and nitrogen, which are adjusted on a breath-by-breath basis to clamp P_ETCO₂ at a consistent threshold. It is recommended that end-tidal forcing systems be used when completing group differences; however, the authors do acknowledge the limitations of this technique, requiring specialized equipment and training. ⁶¹ Further, the end-tidal clamping during NVC and dCA assessments would enhance the internal validity of a study, eliminating the confounding effect of respiratory changes that occur as a byproduct of the NVC and dCA stimuli. ¹⁴

Although numerous benefits of TCD exist, a large concern surrounds the inability to quantify vessel diameter. ^75,86 This means that TCD is only capable of quantifying the velocity of red blood cells moving through the intracranial arteries and does not directly measure cerebral blood flow. According to Poiseuille's law, the flow through a vessel will increase exponentially to the fourth power of the radius. ⁸⁷ Thus, CBv likely underestimates flow when the vessel diameter increases, and overestimates when the diameter decreases. ⁷⁵ Based on the aforementioned signaling of neural NOS and eNOS, ^57,58 changes in diameter are likely to occur during NVC, CVR, and dCA tasks. If these changes are not captured, the between-group cerebrovascular regulation differences may be disguised. Techniques such as transcranial color-coded duplex (TCCD) sonography may have utility in the future to provide an assessment of velocity, diameter, and flow. ⁸⁸ However, this technique is in its relative infancy, and is limited by being unable to provide a parallel angle of insonation, which also limits specificity. ⁸⁹

Overall, SRC is a heterogeneous injury with a variety of injury mechanisms and symptomatic presentations. Therefore, employing large sample sizes and completing multimodal assessments could better address the influence of SRC on cerebral blood flow regulation. Multimodal assessment further increases the ability to understand the relationship among different brain regions and structures. For example, complementing TCD with neuronal-based measures (e.g., electroencephalography) and/or additional hemodynamic measures (e.g., functional magnetic resonance imaging, functional near-infrared spectroscopy) would provide a linkage between neuronal firing and hemodynamic changes. ⁹⁰ Additionally, adding in clinical measures or cognitive tests would allow for an enhanced understanding of the link between physiological and cognitive deficits and the symptoms that occur following SRC.