The effects of a dietary supplement on brain function and structure in Junior A ice hockey players: a prospective randomized trial

Participants

The study was reviewed and approved by the Institutional Review Board. Fifty-four male athletes were recruited from the North American Tier 2 Hockey League (NAHL) and the North American Tier 3 Hockey League (NA3HL). Participants were between 18 and 20 years of age (18.73 ± 0.69) at the time of enrollment, spoke fluent English, and were medically cleared to play ice hockey by the team’s medical staff. All subjects provided written consent. Exclusion criteria included allergy to the Synaquell^TM ingredients (Supplemental Table 1: SynaquellTM Ingredients) and contradictions for the NeuroCatch^® Platform (clinically documented hearing issues, hearing device, aid or implant, or an unhealthy scalp). Participants were asked to refrain from taking other dietary supplements and were excluded if they declared noncompliance. All players underwent a preseason medical evaluation and completed the Sideline Concussion Assessment Tool (SCAT5). Team rosters change during the season; therefore, participants who left their hockey team, lost medical clearance to play or did not comply with study protocols were withdrawn from the study. Athletes added to the roster during the hockey season were enrolled and tested (n=5). All participants were randomized into either the control (CON, n=28) or Synaquell^TM (SQ, n=26) group.

Experimental design

This prospective, randomized clinical trial utilized repeated measures and a placebo group to determine the effects of a combination dietary supplement, Synaquell^TM, on brain function and structure. Pre-season testing included: brain vital signs recorded using the NeuroCatch^® portable evoked potential platform, the K-D rapid number naming test, and venous blood analysis for NfL. Administered clinical neuropsychological questionnaires included: SCAT5, Neuro-Quality of Life Short Form Version 2.0 – Cognitive Function (nQoL), and Patient Health Questionnaire – 4 (PHQ-4). These questionnaires were repeated after the hockey season, along with the Patient Global Impression of Change (PGIC). Post-season data collection was completed with 7 days of the last regular season game.

Brain vital signs

NeuroCatch^® (Version 1.1) incorporates a portable 8-channel g. Nautilus EEG system (Gtec Medical Engineering, Austria). Testing involves a 6-min automated stimulus sequence consisting of auditory tones (oddball stimuli with standard and deviant tones) interspersed with spoken prime-target word pairs (equally semantically congruent or incongruent). Participants were instructed to listen attentively, passively and remain still with their eyes fixed on a cross positioned at eye level 2 m away. Scans were conducted in a quiet, closed room away from distractions. Pre- and post-season scans were conducted in the same facility to maintain a consistent environment.

The three scalp electrodes were referenced to an electrode clipped to the right earlobe. Disposable, Ag/AgCl, adhesive electrodes were used for electro-oculogram recording from the supra-orbital ridge and outer canthus of the left eye. g. GAMMAsys electrode gel was applied to each location to ensure conductivity. Skin-electrode impedances were maintained below the standard 30kΩ threshold at each site.

Scan quality results were evaluated immediately, with all EEG data subsequently processed off-line in Python. Pre-processing was performed using a fourth-order Butterworth filter (0.1–20 Hz) and a notch filter (60 Hz). Adaptive filtering (He et al. 2004) was used to correct for ocular artifacts, which were recorded from electrooculogram locations. ERPs were derived from Fz, Cz, and Pz channels and segmented (−100ms to 900 ms), baseline corrected, and averaged by stimulus condition. Brain vital sign ERP responses were manually verified by a blinded reviewer by identifying the relevant local maxima/minima within expected polarities and temporal ranges for the N100, the P300 and the N400. Peaks were then evaluated for both latency (response time) and amplitude (synchronous processing), for a total of six measures. The N400 amplitude was recorded as the mean voltage from 350-450 ms. N400 latency was recorded as the response time of the selected peak.

King-Devick

K-D testing was completed in a quiet clinical setting both pre-season and post-season. Testing was conducted on an electronic tablet that was held at a comfortable reading distance. Players were instructed to read the numbers aloud as quickly and as accurately as possible from left to right and top to bottom. They completed an untimed demonstration card, followed by three timed test cards. The summed time of the three test cards was the player’s K-D score. At baseline, players completed the test twice and their best score was recorded. At post-season, the players completed the test once and their resultant K-D score was recorded.

Blood Sampling

Non-fasting venous blood samples were collected from the antecubital fossa by standard venipuncture procedures. Samples were collected in Serum Separator Tubes containing a clot activator and serum gel separator. Serum samples were assayed on the same instrument and by the same person in a blinded fashion. NfL concentrations in plasma were measured with the NF-Light digital immunoassay (Quanterix, Cat#103186) using the HD-X Analyzer per the manufacturer’s protocol. In brief, samples were thawed on ice, mixed thoroughly by low-speed vortex and centrifuged at 10,000 g for 5 minutes before transferring samples to 96-well plates. Samples were diluted 1:4 by the instrument and tested in duplicate. In addition to participant serum samples, 8 calibrators and 2 quality control samples provided with the kits were included in the assay. Concentrations were interpolated from the standard curve using a 4-parameter logistic curve fit (1/y2 weighted).

Supplementation

The CON group received a placebo while the SQ group received the supplement, Synaquell^TM. The ingredients for Synaquell^TM and placebo supplements are listed in Supplemental Table 1. Athletes mixed one scoop (7.7 grams) of the provided powder with water and ingested twice daily. Supplements were administered by research team staff at team practices and home games. Supplements were prepared for players on road trips or days without organized team activities.

Research staff tracked all doses administered during practices and home games. Players self-reported all doses taken outside the supervision of the research staff using an internet-based or paper survey. All subjects and research staff responsible for data collection were blinded to the group assignment. Both groups discontinued the intervention after the final regular season hockey game. The NAHL played 60 games and NA3HL played 47 games during this study. Both teams participated in two 60-minute ice hockey games, and 4 or 5 practices lasting 60 to 90 minutes each week.

Statistical Analysis

Statistical analyses were conducted using SPSS (IBM, New York, USA). Descriptive statistics were calculated for all dependent variables. T-tests were conducted to evaluate differences between groups across participant demographics for age, height, weight, number of games played, self-reported concussion history, and intervention compliance (as % of doses taken). The data from this study are available upon request from the corresponding author.

Multivariate Analysis – Objective function and structure assessments

A repeated measures multivariate general linear analysis of variance (GLM) was used to investigate the effect of group allocation on dependent variables (6 brain vital sign measures, K-D, and NfL) as a factor of time (pre- to post- season). Prior concussion reporting and number of games played during the season were included as covariates. The number of prior concussions was derived from subjective participant recall as a standard aspect of the SCAT5 questionnaire, clinically diagnosed concussions during the season were not monitored as part of this study design. Post-hoc tests (both within and between subjects) were conducted to evaluate effects within individual dependent variables.

Multivariate Analysis – Clinical Questionnaires

The statistical model design was then repeated using only the four clinical questionnaires (SCAT5, nQoL, PHQ-4, PGIC) as the dependent variables. PGIC scores for all participants at pre-season was set to 0, given that it is a retrospective analysis of change.

Brain vital signs univariate analysis

To compare with prior evaluations of pre- to post-season changes of brain vital signs in contact athletes, grand average ERP waveforms (means +- 95% confidence intervals) were generated for deviant tone and incongruent word responses for each group. Paired, two-tailed Welch’s t-tests were performed on regions of interest identified on the waveforms as areas of deviation between the confidence intervals. Standardized radar plots were also generated for brain vital signs to indicate relative changes in scores over time.

Participants

Of the 28 players randomized into the CON group, five players withdrew, seven left their team, and one was injured reducing the sample size to N=15. Of the 26 players randomized into the SQ group, three withdrew, six left their team, one was injured, and one dataset was compromised reducing the sample size to N=15. Due to supply chain delays, the intervention began 105 days after pre-season testing of the NA3HL team and 90 days after pre-season testing of the NAHL team. By this time, 26 NAHL games and 25 NA3HL games had occurred. Players were supplemented for the remaining 84 days (22 games) of the NA3HL season and 112 days (35 games) of the NAHL season. Postseason testing occurred 189 days after pre-season testing for the NA3HL team and 202 days for the NAHL team. There were no significant differences between groups across all demographics (Table 1).

Multivariate Analysis – Objective assessments

The multivariate repeated-measures general linear model (Table 2) showed a significant overall within-subjects interaction effect of Group with Time (p = 0.036), which identified differential changes between the CON and SQ groups from pre- to post-season across all dependent variables. Post-hoc tests identified specific significant differences in both N100 latency (p = 0.005) and K-D score (p = 0.043).

There was also a significant, multivariate within-subjects effect between Time and Concussion History (p = 0.001). Post-hoc tests showed that the number of self-reported concussions, independent of group allocation, was significantly related to changes in outcome measures over time for the N100 latency (p = 0.046), N400 latency (p = 0.007), and NfL score (p = 0.012). There were no significant multivariate effects between subjects.

Multivariate Analysis – Clinical Questionnaires

The multivariate repeated-measures general linear model for the clinical questionnaires (Table 3) showed no significant effects of group by time (p = 0.892). There were no significant effects or interaction effects for the covariates at any level of the model.

Violin plots compare the objective measures and clinical questionnaires as a function of pre- versus post-difference for the CON and the SQ groups (Figure 1). The plot enables examination of the distributions of within-subject changes, separated by test type. Best- fit CON and SQ group differences are also line plotted with associated p-values for time effects from the respective multivariate GLM post-hoc tests. The violin plots show distribution differences between the CON and SQ groups in terms of pre- versus post- changes and highlight sensitivity differences between objective and subjective tests.

• Download figure
• Open in new tab

Figure 1:

Violin plots showing pre- versus post- season differences in CON (light blue) and SQ (dark blue) groups. Plots are separated by test type showing objective measures (left side) and subjective measures (right side). Lines connecting plots show best fit differences with statistical difference above (* P < 0.05, ** P < 0.01). Note the statistical differences are from post-hoc tests in the multivariate GLM, which do not show significant effects from other tests (e.g., N400 amplitude reduction in brain vital sign waveform analysis). Examination of the violin plots show distribution differences between CON and SQ, along with sensitivity differences between objective and subjective tests. SCAT-5: The Sport Concussion Assessment Tool 5, PHQ-4: Patient Health Questionnaire-4, nQoL: The Neuro-Quality of Life Short Form Version 2.0 – Cognitive Function, and PGIC: Patient Global Impression of Change questionnaire.

Brain vital signs waveform analysis

Grand average NeuroCatch ERP waveforms (+- 95CI of the mean) for CON (left) and SQ (right) groups were created for both deviant tone and congruent word stimuli (Figure 2). Regions of interest were identified for both the N100 latency, and the N400 amplitude components. Subsequent t-tests at each group identified significant delay in the N100 latency (p = 0.048) in the CON group, but not the SQ group (p = 0.137). Similarly, there were significant reductions in N400 amplitude (p = 0.017) in the CON group, which was not present in the SQ group (p = 0.506). Standardized brain vital sign radar plots compare CON (left) and SQ (right) with pre- versus post-season comparisons for both groups (Figure 3).

• Download figure
• Open in new tab

Figure 2:

Grand average NeuroCatch ERP waveforms (+- 95CI of the mean) for CON (left) and SQ (right) groups (*p <0.05). Pre-season is in blue and post-season is in orange. The N100 and P300 to deviant tone stimuli (top row) and the N400 to incongruent word stimuli (bottom row). Time is on the x-axis (ms) and voltage is on the y-axis (µVs). Note the N100 latency delay and the N400 amplitude reduction in the CON group that is not present in the SQ group.

• Download figure
• Open in new tab

Figure 3:

Standardized NeuroCatch^® radar plots showing brain vital sign results for each group at each time point (*p < 0.05). All 6 brain vital sign measures are plotted for both the CON (left) and SQ (right) groups. Pre-season results are in blue and post-season results are in orange. The dashed line denotes the median from prior normative data. Note the CON results replicate prior subconcussion impairment findings (Fickling et al., 2019, 2021a,b).

Main findings

The aim of this study was to investigate the effects of the dietary supplement Synaquell^TM on brain function and structure. Multivariate analysis, including brain vital signs, K-D, NfL, showed significant within-subject differences between groups in changes from pre-to post-season (p = 0.036) (Table 2, Figure 1; Primary Hypothesis). Post-hoc tests revealed significant differential changes between groups in functional variables, including N100 latency and King-Devick scores. Specifically, the N100 latency became slower in the CON group and faster in the SQ group from pre-season to post-season (Figure 1, Supplemental Table 2). While K-D scores in both groups improved from pre-season to post-season, the improvement in the CON group was faster (Figure 1, Supplemental Table 2).

Additionally, a significant within-subject multivariate effect (p = 0.001) was observed for Concussion History * Time. Post-hoc testing identified significant changes in functional measures of N100 latency, N400 latency, and the structural NfL measure. This suggests that increased head impact exposure is linked to greater changes in both brain function and structure over the course of the season, regardless of group allocation.

The secondary analysis replicated previous findings on neurocognitive changes in athletes with no intervention (i.e., the CON group), specifically the delay in the N100 latency and decrease in N400 amplitude (Figure 3, Secondary Hypothesis). In contrast, no significant change was observed in the SQ group from pre- to post-season (Figs. 2&3).

Differential changes in objective brain function measures between the CON and SQ subjects from pre- to post-season supports the use of the dietary supplement Synaquell^TM, to support brain function and structure in healthy, male athletes. To our knowledge, this is the first human study in a population of healthy, non-concussed athletes in which a multi-ingredient dietary supplement was found to be effective in supporting neurocognitive function and brain structure (Feinberg et al. 2023). The most similarly advanced studies in such a population have investigated Omega-3 Fatty Acids in relation to brain structure (Oliver et al. 2016; Heileson et al 2021; Mullins et al. 2022). Our results coincide with the findings of Oliver et al. and Heileson et al., in which serum NfL was attenuated by the daily administration of DHA in athletes (Oliver et al. 2016; Heileson et al. 2021). Mullins et al. 2022 alternatively found that daily DHA supplementation did not reduce serum NfL levels over the course of the season (Mullins et al. 2022). Studies examining the effects of single-ingredient dietary supplementation on neurocognition and brain structure have been conducted in concussed populations post-injury (Feinberg et al. 2023). Multiple ingredients have the potential to impact a variety of pathways and processes in the neurometabolic cascade, which may improve effectiveness (Lucke-Wold et al. 2016).

Brain vital signs

The current study provides the latest evidence that head impact exposure affects cognitive processing (N400) and auditory sensory processing (N100). A pattern involving the N100 and N400 is emerging across different ages and sports. Specific responses vary slightly in terms of amplitude, latency and waveform shape across studies, but the consistency of N100 and N400 impairment in athletes exposed to repetitive head impacts is notable (Fickling et al. 2019, 2021a,b).

While the P300 component has shown little change to-date when examining group average data, the results of the multivariate analyses indicate greater complexity with respect to individual variability and factors affecting head impact exposure. For instance, the P300 amplitude and latency, as markers of basic attention, co-varied significantly with subjective concussion history (as did N400 latency). Similarly, significant covariance occurred between the N100 and the K-D test for pre- versus post- comparisons by group by games played (as did the N400 for pre- versus post by games played). Of interest, NfL results significantly co-varied in terms of pre- versus post- season and subjective concussion history. Taken together, the current findings suggest that while N100 and N400 changes may serve as an initial indicator of subconcussive impairment, additional brain vital sign features may also be informative as objective markers for underlying sub-concussive impairment profiles across a spectrum of differential head-impact exposure between players.

Objective assessments of function and structure

There were no significant differences between the groups in clinical neuropsychological questionnaires pre- to post-season despite significant differences in objective neurocognitive measures. This observation is not surprising, given that clinical questionnaires such as the SCAT5 are subjective in nature, rely heavily upon symptom reporting, and healthy, non-concussed athletes that are exposed to repetitive head impacts typically do not show observable physical signs of neurological impairment (Lember et al. 2021). The SCAT5 has been shown to differentiate concussed and non- concussed athletes in an acute concussion diagnosis setting; however, it is less useful in monitoring recovery (Echmendia et al. 2017). The other neuropsychological questionnaires (PHQ-4, NQoL, and the PGIC) also did not show differences between groups that were detected by objective measures. Nonetheless, these questionnaires provide a subjective, yet insightful glimpse into the functional state of the individual’s brain health.

The current findings highlight the importance of expanding upon studies investigating the relationship between dietary supplementation, cognitive function and brain structure in athletes exposed to repetitive head impacts. Also important to note is the relationship between objective measures of concussion and subjective clinical symptoms. Batteries of objective tools such as the brain vital signs, K-D, and NfL may be able to detect neurocognitive impairments that are at or below current diagnostic thresholds for concussion. A critical issue involves the point at which cognitive impairments manifest as detectable clinical-behavioral symptoms. Further investigation is needed to investigate if dietary supplements alter or prevent the emergence of clinical-behavioral symptomatology.

Limitations

This interventional trial has several limitations. First, the cohort was exclusively male Junior-A hockey players; therefore, further studies are required to generalize the results. These investigations should also examine factors related to concussion diagnosis, including gender, age, and mechanism of injury (e.g., blast injuries). Second, variables related to the specific dietary supplement, Synaquell^TM, need further analysis. Synaquell^TM doses were not standardized by body weight, and individuals with higher body mass may require higher dosage to experience the same effects on neurocognition. Future studies should evaluate optimal dosages, ingredients, and the potential for individually tailored dietary supplementation. Third, supplementation did not begin at the beginning of the season due to supply chain challenges. The athletes participated in a portion of the season without receiving the supplement. Season long supplementation may have contributed to greater neurocognitive changes (or lack thereof) in the SQ group. Fourth, our study examined only pre- versus post- timelines, and further time points should be investigated in future studies, including whether dietary supplements support the brain during recovery from a clinically diagnosed concussion. Finally, the current study did not include quantified impact evaluation, so the relationship with impact exposure remains to be determined.