Exercise Intensity and Recovery on Circulating

Brain-derived Neurotrophic Factor

JOSHUA T. REYCRAFT1, HASHIM ISLAM2,3, LOGAN K. TOWNSEND2,4, GRANT C. HAYWARD1,
TOM J. HAZELL2, and REBECCA E. K. MACPHERSON1
Downloaded from http://journals.lww.com/acsm-msse by BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsIHo4XMi0hCyw

1
Department of Health Sciences, Brock University, St. Catharines, ON, CANADA; 2Department of Kinesiology and Physical
Education, Wilfrid Laurier University, Waterloo, ON, CANADA; 3School of Kinesiology and Health Studies, Queen’s University,
Kingston ON, CANADA; and 4Department of Human Health and Nutritional Sciences, University of Guelph, Guelph, Ontario, CANADA
CX1AWnYQp/IlQrHD3i3D0OdRyi7TvSFl4Cf3VC1y0abggQZXdgGj2MwlZLeI= on 05/26/2023

ABSTRACT
REYCRAFT, J. T., H. ISLAM, L. K. TOWNSEND, G. C. HAYWARD, T. J. HAZELL, and R. E. K. MACPHERSON. Exercise Intensity
and Recovery on Circulating Brain-derived Neurotrophic Factor. Med. Sci. Sports Exerc., Vol. 52, No. 5, pp. 1210–1217, 2020. Introduction:
Brain-derived neurotrophic factor (BDNF) is an exercise-induced neurotropin mediating neuroprotection and synaptic plasticity. Although exer-
cise intensity is implicated as a potentially important mediator of BNDF release after exercise, the optimal exercise stimulus (interval vs contin-
uous) and intensity (submaximal vs supramaximal) for augmenting circulating BDNF levels remains unknown. Irisin, an exercise-driven
myokine, may also contribute to neuroprotection by upregulating BDNF. Purpose: To examine the response and recovery of plasma BDNF
and irisin after acute exercise of differing intensities. Methods: Eight males (23.1 ± 3.0 yr of age; V̇O2max 51.2 ± 4.4 mL·kg−1·min−1) completed
four acute exercise sessions: 1) moderate-intensity continuous training (MICT, 65% V̇O2max); 2) vigorous-intensity continuous training (VICT,
85% V̇O2max); 3) sprint interval training (SIT, “all out”); and 4) no exercise (CTRL). Blood was collected preexercise as well as immediately,
30 min, and 90 min postexercise. Plasma BDNF and irisin were assessed with commercially available enzyme-linked immunosorbent assay
kits. Results: Plasma BDNF levels increased immediately after exercise in the SIT group (P < 0.0001) with plasma concentrations recovering
30 and 90 min postexercise. The BDNF levels after MICT were reduced 30 min postexercise compared with immediately postexercise
(P = 0.0189), with no other changes across time points in MICT and VICT groups. Plasma BDNF area under the curve in SIT was signifi-
cantly higher compared with CTRL, MICT, and VICT (P = 0.0020). No changes in plasma irisin across exercise groups and time points were
found (P > 0.9999). Conclusions: Plasma BDNF levels increased in an intensity-dependent manner with SIT eliciting the highest BDNF concentration
immediately postexercise. These results identify SIT as a time-efficient exercise modality to promote brain health through BDNF release. Key Words:
SPRINT INTERVAL TRAINING, HIGH-INTENSITY INTERVAL TRAINING, AEROBIC EXERCISE, BDNF, LACTATE, IRISIN

C
haracterized by the progressive deterioration of cogni- to sustain cognition during aging would have enormous soci-
tive and functional ability beyond that of normal ag- etal and economic impact.
ing, dementia is now the seventh leading cause of The transient short-term functional improvements and long-
death worldwide with Alzheimer’s disease (AD) making up term neuroprotective benefits of exercise have been attributed
60% to 70% of cases (1). Physical inactivity shows a greater to the activity-induced expression and release of the neurotrophin
population attributable risk for AD development compared family member brain-derived neurotrophic factor (BDNF) (3). At
with midlife obesity and other related morbidities combined a neuronal level, BDNF release promotes synaptogenesis which
(hypertension, diabetes mellitus) (2). Although exercise in turn directly mediates cognitive processes, such as memory,
training is an effective treatment method for obesity and learning, and emotion (4,5). Interestingly, BDNF is able to cross
metabolic dysfunction, the augmented effect size of physi- the blood–brain barrier (BBB) and cross-species studies show
cal inactivity in AD risk reveals a direct mechanism by that whole blood BDNF concentrations reflect BDNF levels in
which exercise offers neuroprotective effects. Given the the brain (6,7). Further, it is known that basal levels of plasma
health benefits of physical activity and its wide accessibility BDNF decline with age (8) and that reduced circulating BDNF
APPLIED SCIENCES

at any age, determining how exercise can be used as medicine has been demonstrated in patients with major depression, obe-
sity, and type 2 diabetes (9,10). In addition, reduced circulating
BDNF has been implicated with a faster cognitive decline in pa-
Address for correspondence: Rebecca EK MacPherson, Ph.D., Department of
Health Sciences Brock University St. Catharines, Ontario, Canada; E-mail: tients with AD compared with patients with a slower cognitive
rmacpherson@brocku.ca. decline (11). The prominence of reduced plasma BDNF in AD
Submitted for publication July 2019. and populations at greater risk for AD (e.g., obesity, type 2 di-
Accepted for publication December 2019. abetes mellitus, aging) demonstrate the importance of examin-
0195-9131/20/5205-1210/0 ing interventions that increase BDNF and preserve brain health.
MEDICINE & SCIENCE IN SPORTS & EXERCISE® Aerobic exercise training is known to elevate resting plasma
Copyright © 2019 by the American College of Sports Medicine BDNF (12,13), demonstrating the potential of regular exercise
DOI: 10.1249/MSS.0000000000002242 to increase circulating BDNF. Acutely, it has been reported

1210

Copyright © 2020 by the American College of Sports Medicine. Unauthorized reproduction of this article is prohibited.
 that the brain contributes 70% to 80% of the circulating BDNF Laurier University for a previously published study (25). Par-
postexercise (14). This increase in circulating BDNF postexer- ticipants were nonsmokers and determined healthy through
cise has been demonstrated across ages, sex, and within patient the Physical Activity Readiness Questionnaire (PAR-Q+)
populations (panic disorder, AD, major depressive disorder) health questionnaire before study enrolment. All participants
(15–18). A recent meta-analysis found that there is a relation- were physically active (less than or equal to three sessions per
ship between exercise duration and BDNF in serum and week) and not involved in any systemic training program for
plasma of healthy adults; however, the majority of evidence at least 6 months before study participation. Participants were
points toward a stronger intensity-dependent relationship, with instructed to refrain from any dietary supplements at the time
Downloaded from http://journals.lww.com/acsm-msse by BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsIHo4XMi0hCyw

higher intensities provoking greater circulating BDNF concentra- of the study and refrain from any physical activity, alcohol, or
tions (15,19–22). Yet, the optimal exercise intensity to provoke caffeine for >24 h before experimental sessions. Experimental
the highest and longest lasting circulating BDNF remains un- details were fully explained to all participants and all provided
CX1AWnYQp/IlQrHD3i3D0OdRyi7TvSFl4Cf3VC1y0abggQZXdgGj2MwlZLeI= on 05/26/2023

known. Given the relationship between BDNF release and exer- written informed consent before any data collection. This study
cise intensity, sprint interval training (SIT) may provide a time was approved by the Research Ethics Board at Wilfrid Laurier
efficient exercise that optimally increases BDNF. University in accordance with the 1964 Declaration of Helsinki.
In addition to the peak BDNF elicited by exercise, it is im- Experimental design. The experimental design has been
portant to establish which exercise will elicit the longest in- described previously (25). Briefly, each participant completed
crease in BDNF postexercise. Studies examining the BDNF four different experimental sessions, with each session taking
recovery profile after high-intensity acute exercise have re- place >1 wk apart using a systemic rotational order using a
ported inconsistent results. For example, after the completion counterbalanced Latin square design. Experimental interven-
of a graded exercise test to volitional fatigue serum and plasma tions consisted of a resting control session (CTRL; no exercise)
BDNF recovered by 30 and 60 min, respectively (23). How- in addition to three running-based exercise protocols:
ever, a running-based hight intensity interval training (HIIT) 1) moderate-intensity continuous training (MICT) (65% of
protocol at maximal intensity (100% V̇O2peak) demonstrated maximal oxygen uptake [V̇O2max]); 2) vigorous-intensity con-
serum BDNF recovery within 60 min, though no earlier blood tinuous training (VICT) (85% V̇O2max); and 3) SIT (four 30-s
draws were collected (22). As studies to date have neither bouts of “all-out” running interspersed with 4-min rest periods).
demonstrated intensity-dependent BDNF response in plasma Blood draws were obtained at four time-points throughout each
nor postexercise BDNF changes after SIT, the effect of exer- 3-h experimental session. Blood draws for the immediately
cise intensity and SIT on plasma BDNF recovery is unknown. postexercise samples were taken within 30 s of completing
Interestingly, recent in vivo rodent and in vitro cell cul- the 5-min cool-down for all participants. Participant’s energy
ture studies have suggested a novel mechanism by which intake was recorded over a 3-d period and participants were
skeletal muscle contraction secretes myokines into the cir- asked to replicate their dietary intake for 24 h before each exper-
culation, which may readily cross the BBB and induce cen- imental session.
tral Bdnf expression. One potential candidate is circulating Preexperimental procedures. One week before the
irisin and its membrane-bound form fibronectin type III first experimental session, all participants completed a famil-
domain-containing protein 5 (FNDC5) (24). Mediated iarization session in which the participants were introduced
through peroxisome proliferator-activated receptor gamma to testing procedures and equipment. During the familiariza-
coactivator 1-alpha (PGC-1α), in vivo and in vitro models tion session, participants’ V̇O2max was determined through a
show Bdnf expression to be dependent on FNDC5 (24). Fur- graded exercise test to exhaustion on a motorized treadmill
thermore, adenoviral vector administration of FNDC5 into (4Front, Woodway, WI) as described previously (25). Oxygen
the liver of mice transiently increased circulating irisin and en- consumption and carbon dioxide production (V̇O2 and VCO2,
hanced hippocampal Bdnf expression (24). respectively) were measured continuously using an online
This study tested the hypothesis that exercise intensity is an breath-by-breath gas collection system calibrated using known
important mediator of changes in plasma BDNF levels in gas concentrations and a 3-L syringe for flow. To ensure com-
healthy male adults. Specifically, this study aimed to examine: fort and accurate gas measurement recordings, participants
1) the intensity-dependent plasma BDNF response and recov- wore a fitted silicon facemask. Participant’s HR was recorded
APPLIED SCIENCES
ery profile following different exercise protocols across an in- beat-to-beat using an integrated HR monitor.
tensity continuum; and 2) investigate exercise-induced changes Determination of V̇O2max consisted of a 5-min treadmill
in circulating irisin as a candidate modulator of BDNF expres- warm-up followed by participants running at a self-selected
sion in response to running-based exercise. Results from this pace (5–7 mph) maintained throughout the test. Incremental
investigation will help to identify the optimal exercise inten- increases in workload (2% grade) were applied every 2 min
sity for promoting brain health. until volitional fatigue was achieved. V̇O2max was defined by
as the greatest 30-s average at which V̇O2 values plateaued
(increase <1.35 mL·kg−1·min−1) despite increases in work-
METHODS load, or two of the following criteria: 1) RER value >1.10;
Participants. Eight active young males participated in this 2) maximal HR (within 10 bpm of age-predicted maximum
study and all experimental sessions were conducted at Wilfrid [220 age]) and/or; 3) voluntary exhaustion. All participants

SPRINT INTERVAL TRAINING INCREASES BDNF Medicine & Science in Sports & Exercise® 1211

Copyright © 2020 by the American College of Sports Medicine. Unauthorized reproduction of this article is prohibited.
 met the RER and HR criteria used for the determination of subsequent plasma-parameter analysis. To measure blood lac-
V̇O2max, whereas six of the eight participants also achieved a tate, a blood droplet was collected from one of the Vacutainer
plateau in V̇O2 values. After a short 5-min treadmill cool- tubes and placed on a lactate strip using a hand-held analyzer
down and 20-min rest period, the running speed/grade required (Accutrend Lactate; Roche Diagnostics, Mannheim, Germany).
to elicit the appropriate workloads for MICT (65% V̇O2max) and Commercially available enzyme-linked immunosorbent assay
VICT (85% V̇O2max) were determined. Participants began by kits were used to determine plasma concentrations of BDNF
jogging at moderate pace (5 mph) with incremental increases in (Cat. CYT306; EMD Millipore BDNF enzyme-linked im-
speed (1–2 mph) applied every 3 min until the speed correspond- munosorbent assay [ELISA]) and Irisin (Cat. EK-067-29;
Downloaded from http://journals.lww.com/acsm-msse by BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsIHo4XMi0hCyw

ing to the appropriate workload was achieved and recorded. Phoenix Pharmaceuticals Inc.). The BDNF ELISA kit was
Experimental sessions. Participants arrived at the labo- specific to human BDNF with no significant cross-reactivity
ratory at 8:00 AM after an overnight fast and the experimental with other members of the neurotrophin family (NGF, NT
CX1AWnYQp/IlQrHD3i3D0OdRyi7TvSFl4Cf3VC1y0abggQZXdgGj2MwlZLeI= on 05/26/2023

session was completed over a 3-h period (Fig. 1). Once arrived, 4/5, or NT3). All samples were run in duplicate.
participants were given a standardized test meal consisting of a Statistical analysis. All participant data were analyzed
Chocolate Chip Clif Bar (7 kcal·kg−1 body mass) and water, using GraphPad Prism Version 6 (La Jolla, CA). To account
which was provided ad libitum throughout the session. Partic- for individual variability within participants in absolute protein
ipants were given 15 min to consume the test meal, followed by concentrations, protein concentration changes at each time point
a 30-min sitting rest. After the rest, exercise commenced at were expressed relative to the group’s average protein concentra-
8:50 AM consisting of a 5-min standardized warm-up (3.5 mph) tion at baseline. Area under the curve (AUC) for plasma BDNF
on a motorized treadmill, followed by a 30-min running based and Irisin fluctuations across the experiment session was cal-
exercise protocol (MICT or VICT), and a 5-min standardized culated using the change in each variable relative to baseline.
cool-down. To match exercise protocol duration between exer- One-way repeated measures ANOVA was used to compare
cise groups, SIT sessions were allocated an additional 12 min of plasma BDNF and irisin AUC values across sessions. Two-
rest before warm-up, followed by the required 18 min for the way repeated-measures ANOVA (session–time) was used to
SIT exercise protocol (four 30-s “all out” running efforts inter- compare differences in plasma BDNF and irisin between ex-
spersed with 4-min recovery). Participants in the SIT group ran perimental sessions at each time point. Post hoc analysis using
on a specialized self-propelled treadmill (HiTrainer Pro, QC, Tukey’s multiple comparisons test was used when necessary.
Canada). Gas exchange and HR were measured continuously To estimate effect sizes for between-protocol differences
throughout the exercise component of the experimental ses- in changes in BDNF, Cohen’s d calculation was used. Sig-
sion. Upon completing the running-based exercise protocol, nificance was set at P < 0.05. All data are presented as
participants remained in the laboratory until 11:00 AM for an means ± standard error.
additional 90-min rest period. The experimental protocol within
the CTRL sessions was similar, with time allocated toward
running-based protocol (8:50 AM to 9:30 AM) replaced with
RESULTS
quiet rest. Venous blood samples were collected from partici- Participant characteristics. As previously published
pants at 8:45 AM (preexercise), 9:30 AM (immediately postex- (25), participants were male (n = 8), 23.1 ± 3.0 yr of age, and
ercise), 10:00 AM (30-min postexercise), and 11:00 AM (90-min had a mean V̇O2max of 51.2 ± 4.4 mL·kg−1·min−1
postexercise). Specifically, blood samples collected immedi- (4.1 ± 0.27 L·min−1). Additionally, participants had a mean
ately postexercise were drawn from participants within 30 s of height of 178.2 ± 2.7 cm; weight of 78.7 ± 8.1 kg; and body
cessation of the standardized cool-down. mass index of 24.8 ± 2.3 kg·m−2 (Table 1).
Blood processing and biochemical analysis. All blood Brain-derived neurotrophic factor. Plasma BDNF
samples were collected by venipuncture from the antecubital vein was determined across four exercise groups (CTRL, MICT,
while lying in a supine position. Two samples of blood (3 mL VICT, SIT), and four time points (PRE, POST, 30-POST, and
of whole blood each) were collected into separate prechilled 90-POST). At baseline, there were no differences (P > 0.19)
Vacutainer tubes coated with K2 EDTA (5.4 mg) at each time in absolute BDNF plasma concentrations between groups
point. Tubes were inverted 10 times, centrifuged at 3000g for (CTRL: 4934 ± 1048 pg·mL−1; MICT: 4775 ± 1095 pg·mL−1;
APPLIED SCIENCES

10 min, and the plasma supernatant was then dispended into VICT: 5470 ± 1535 pg·mL−1; SIT: 3168 ± 420 pg·mL−1). Im-
Eppendorf tubes. Plasma samples were stored at −80°C until mediately postexercise, the SIT group had significant increases

FIGURE 1—Timeline of 3-h experimental exercise session.

1212 Official Journal of the American College of Sports Medicine http://www.acsm-msse.org

Copyright © 2020 by the American College of Sports Medicine. Unauthorized reproduction of this article is prohibited.
 TABLE 1. Participant characteristics. respectively; Fig. 2C) and CTRL levels (P = 0.0002 and
Age 23.1 ± 3.0 yr P = 0.0053, respectively). MICT postexercise plasma BDNF
Mean V̇O2max 51.2 ± 4.4 mL·kg−1·min−1 (4.1 ± 0.3 L·min−1)
Height 178.2 ± 2.7 cm
was no different to preexercise levels and CTRL (P = 0.0567
Weight 78.7 ± 8.1 kg and P = 0.0713, respectively). Lastly, we examined the changes
BMI 24.8 ± 2.3 kg·m−2 in plasma BDNF concentration immediately after exercise ses-
Values are expressed as means±SD. BMI, body mass index. sion relative to baseline. Plasma BDNF significantly increased
postexercise in the SIT group compared to CTRL (P =
0.0277), with no other changes observed in continuous exercise
Downloaded from http://journals.lww.com/acsm-msse by BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsIHo4XMi0hCyw

in BDNF compared to CTRL (P = 0.0002; d = 1.24), MICT groups (MICT and VICT) compared to CTRL (P = 0.2030 and
(P = 0.0134; d = 1.12), and VICT (P = 0.0120; d = 1.23) P = 0.2255, respectively; Fig. 2D).
(Fig. 2A). Within the SIT group, BDNF levels increased imme- Irisin. In addition to BDNF, plasma irisin levels were de-
CX1AWnYQp/IlQrHD3i3D0OdRyi7TvSFl4Cf3VC1y0abggQZXdgGj2MwlZLeI= on 05/26/2023

diately postexercise compared to preexercise (P < 0.0001). termined across exercise groups and time points. At baseline,
BDNF returned to levels similar to baseline at 30-min there was no difference between any of the groups (P > 0.9999;
(P = 0.1481) and 90-min postexercise (P = 0.7222). Further- Fig. 3A). Further, no significant differences in CTRL, MICT,
more, a significant reduction in BDNF was observed within and SIT groups, at any time points, or when comparing groups
the MICT group at 30-min postexercise compared to immedi- within the same time point, were observed (P > 0.05). Irisin
ately postexercise (P = 0.0189; Fig. 2A). No other changes in the VICT group however, did increase 30-min postexer-
were observed. cise compared to immediately postexercise (P = 0.0357). There
When examining the AUC, plasma BDNF levels were in- were also no significant changes in plasma irisin AUC levels
creased across the experimental session in the SIT group com- (P = 0.1732; Fig. 3B), irisin from preexercise to immediately
pared to CTRL (P = 0.0020), MICT (P = 0.0477), and VICT postexercise (P = 0.4026; Fig. 3C), and irisin immediately after
(P = 0.0279). No significant differences were revealed be- exercise compared to baseline levels (P = 0.1167; Fig. 3D).
tween continuous exercise groups (MICT and VICT) and Correlational results. As recent research demonstrates
CTRL (P = 0.7877; Fig. 2B). Changes in plasma BDNF con- elevated blood lactate levels induce brain BDNF expression
centrations from preexercise to immediately postexercise were in exercise in vivo, correlational analysis was completed be-
also analyzed. Both SIT and VICT increased immediately after tween blood lactate and plasma BDNF. Correlational analysis
exercise compared with preexerise (P = 0.0001 and P = 0.0274, between the peak plasma BDNF immediately postexercise

APPLIED SCIENCES

FIGURE 2—Changes in plasma BDNF concentration across exercise sessions and blood draw time points. A, Changes in plasma BDNF at time points in
each experimental session relative to baseline. a,b,cSIT significantly different from POST CTRL, VICT, and MICT, respectively, P < 0.05. #SIT POST sig-
nificantly different from SIT PRE, 30-POST, and 90-POST (P < 0.01). B, Plasma BDNF AUC across all time points in each experimental session relative to
baseline. C, Changes in absolute plasma BDNF concentration between preexercise to immediately postexercise. D, Change in plasma BDNF concentration
immediately after exercise session relative to baseline. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. All data are presented as means ± standard error.

SPRINT INTERVAL TRAINING INCREASES BDNF Medicine & Science in Sports & Exercise® 1213

Copyright © 2020 by the American College of Sports Medicine. Unauthorized reproduction of this article is prohibited.
 Downloaded from http://journals.lww.com/acsm-msse by BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsIHo4XMi0hCyw
CX1AWnYQp/IlQrHD3i3D0OdRyi7TvSFl4Cf3VC1y0abggQZXdgGj2MwlZLeI= on 05/26/2023

FIGURE 3—Changes in plasma irisin concentration across each exercise group and blood draw time points. (A) Changes in plasma irisin across time points
in each experimental session relative to baseline (n = 8). a30-POST VICT is significantly greater than POST. (B) Plasma irisin AUC across all time points in
each experimental session relative to baseline, no changes observed. (C) Absolute plasma irisin concentrations preexercise and postexercise in each exper-
imental session. No changes were observed. (D) Change in plasma irisin concentration immediately after exercise session relative to baseline. No changes
were observed. All data are presented as means ± standard error.

(~5.5 min after the intervention, including the 5-min cool consisting of four 30-s bouts of all-out sprinting on a manual
down) and lactate as well as irisin were completed. Past liter- treadmill elicited a significantly higher postexercise plasma
ature demonstrates a peak in blood lactate 6 min after exercise BDNF response compared to both 30 min of moderate- and
cessation (26), therefore it is likely that peak lactate levels vigorous-intensity continuous exercise. These findings are in
were obtained after SIT, however this may not be the case line with the intensity-dependent relationship observed in
for MICT or VICT. For this original analysis, we used pre- exercise-induced increase in serum BDNF. After aerobic exer-
viously reported lactate results from Islam et al. (25). A sig- cise where high-intensity endurance exercise at a V̇O2 10%
nificant weak/moderate positive correlation was discovered over the ventilatory threshold (VT) or where exercise at
between the changes in lactate and BDNF (r = 0.3906, ~75% V̇O2max increased postexercise serum BDNF compared
P = 0.03; Fig. 4A). with low-intensity exercise which was 20% below the ventila-
tory threshold or ~45% V̇O2max of the same duration (20).
Moreover, a study utilizing a 20-min HIIT protocol (1-min cy-
DISCUSSION cling at 90% work rate max (Wmax) interspersed with 1-min
The present study provides novel results demonstrating SIT periods of active rest) demonstrated a slightly greater postexercise
to be more effective in elevating plasma BDNF compared with
moderate- and vigorous-intensity continuous exercise modali-
ties. We further demonstrate that the elevated postexercise
plasma BDNF profile was short-lived, recovering within
30 min of exercise cessation. Interestingly, although no
APPLIED SCIENCES

changes in plasma irisin were observed in response to exercise,

a significant positive correlation between blood lactate and
BDNF was found, further supporting the relationship between
exercise intensity and BDNF. These results provide valuable
information in the design of running-based exercise pre-
scription for brain health through provoking a neuroprotec-
tive BDNF response.
In the present study, we demonstrate elevated plasma BDNF
after 30 min of continuous running at vigorous-intensity (85% FIGURE 4—Correlational results between change in plasma BDNF
(pg·mL−1) concentrations and (A) change in lactate (mmol·L−1), (B)
V̇O2max), however only an upward trend after the same duration change in plasma irisin (ng·mL−1). Significance set to P < 0.05. All data
of moderate-intensity (65% V̇O2max) running. Moreover, SIT are presented as means ± standard error.

1214 Official Journal of the American College of Sports Medicine http://www.acsm-msse.org

Copyright © 2020 by the American College of Sports Medicine. Unauthorized reproduction of this article is prohibited.
 serum BDNF response compared with 20 min of high-intensity changes in plasma BDNF postexercise and blood lactate. In
continuous cycling at 70% Wmax (21). In the present study addition to the exercise intensity related relationship between
SIT, designed to reach intensities equal to or greater than BDNF and lactate, recent work has demonstrated that acute
V̇O2max/peak, elicited a robust increase in plasma BDNF com- exercise in vivo generated a lactate-induced hippocampal Bdnf
pared with vigorous-intensity continuous running. These results expression (31). Specifically, augmented lactate levels in the
demonstrate that exercise protocols eliciting maximal effort/ brain during exercise elevate intraneuronal NAD+ levels,
intensity result in the largest BDNF response. which in turn activate SIRT1 and induce the PGC1α pathway
It is important to note the postexercise plasma BDNF recov- promoting FNDC5/irisin, which has been shown to induce
Downloaded from http://journals.lww.com/acsm-msse by BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsIHo4XMi0hCyw

ery profile in investigating the optimal exercise modality to fa- Bdnf expression (31,32). Indeed, the BBB is permeable to lac-
cilitate a BDNF response. Our results demonstrate a rapid tate and exercise further up regulates expression of lactate
recovery of postexercise elevated plasma BDNF, which re- transporter MCT1 at the BBB (33). Moreover, treatment of
CX1AWnYQp/IlQrHD3i3D0OdRyi7TvSFl4Cf3VC1y0abggQZXdgGj2MwlZLeI= on 05/26/2023

turned almost to baseline levels by 30 min after exercise cessa- exercised mice with an MCT1 inhibitor rendering the BBB
tion. As BDNF is readily able to cross the BBB in a impermeable to lactate abolishes exercise-induced Bdnf ex-
bidirectional manner and peripheral levels in the blood reflect pression (31). Therefore, we propose that lactate induced by
that of the brain, exercise modalities that elicit a longer dura- high-intensity exercise, such as SIT, may be a driving factor
tion of elevated BDNF in circulation would have prolonged in stimulating brain Bdnf expression which over time may in-
effects in the brain (6,7). Interestingly, time required for crease neuronal BDNF content for release with subsequent ex-
BDNF recovery after acute exercise has been shown to be de- ercise bouts. One study has previously demonstrated a
pendent on the blood component analyzed. Specifically, se- correlation between serum BDNF and lactate; however, this
rum BDNF was shown to recovery more quickly as is the first study to show a significant positive correlation be-
compared with plasma BDNF (30 min vs 60 min) after the tween change in lactate and BDNF in the plasma (20).
completion of an incremental graded exercise test (23). The Irisin’s role as a peripheral messenger in organ crosstalk has
present study is the first to examine plasma BDNF recovery af- been criticized due to ambiguity regarding the appropriate
ter SIT. Given the short plasmatic BDNF recovery profile, our measure of the biologically functional FNDC5 cleavage prod-
results suggest that SIT does not offer any beneficial effect in uct (34). In fact, studies reporting irisin levels in circulation
maintaining postexercise augmented BDNF, nor might post- have reported a large range from 24 pg·mL−1 to 2e+6 pg·mL−1
exercise plasma BDNF have a prolonged recovery and more and a normal range of irisin in circulation has yet to be estab-
closely align with that of serum. lished (35). Although three irisin ELISA kits have been vali-
Irisin, a muscle-derived myokine has recently been shown dated as accurate and reliable in measuring concentration
to be released by skeletal muscle in exercise and up-regulate changes, reported irisin concentrations are still largely depen-
central BDNF in an endocrine manner (24,27). Therefore, a dent on the commercially available kit used (35). Of the three,
second aim of this study was to examine the effects of exercise the kit by Phoenix Pharmaceuticals used in the present study is
intensity and recovery on plasma irisin. Given the myokines’ considered the current gold standard, with levels consistently
role in increasing energy expenditure and thus a potential tar- falling in the linear range of the curve. Our reported values of
get for treatment in obesity, a number of studies have investi- plasma irisin ranging from 11.44 to 40.95 ng·mL−1 align with
gated and found increased circulating irisin in response to recent studies utilizing the same ELISA kit and showing no
acute exercise in a wide range of modalities (28), although this exercise-induced irisin change in the plasma (36,37). Given that
myokine is highly contentious with other articles being unable irisin levels are equated between plasma and serum, it is unlikely
to detect any effect of acute exercise on circulating levels. Our our results were affected by portion of blood analyzed (38).
results show no change in plasma irisin in response to running- In interpreting the results of the correlational analysis be-
based exercise and no correlation with BDNF. In contrast, tween blood lactate and plasma BDNF after exercise, it is im-
studies have demonstrated both 45 min of MICT (70% portant to note that measurements of the candidate peripheral
V̇O2max) and 16 min of HIIT (4  4 min at 95% HRmax sepa- modulators immediately postexercise (5.5 min after the partic-
rated by 5  4 min active recovery) elicit an increase in plasma ular intervention) were used. Previous work has demonstrated
irisin in healthy adults (29,30). Because this is the first study to that blood lactate concentrations increase approximately 6 min
APPLIED SCIENCES
investigate both irisin and BDNF levels after acute exercise in post–high-intensity exercise and demonstrate no difference
healthy adults, and both have been shown to increase in re- between the fourth, fifth, and sixth minute after exercise cessa-
sponse to acute exercise, further studies examining a corr- tion (26), therefore, we believe that the timing of our postexer-
elation between irisin and BDNF after acute exercise are cise blood lactate measurement (~5.5 min after completion of
required. However, given the lack of response in circulating the last sprint) was a valid approximation of blood lactate re-
irisin in the present study, it is unlikely that elevated plasma sponses after SIT. However, we acknowledge that this timing
BDNF and exercise-induced BDNF content as in the present may not have been optimal for the measurement of blood lac-
study can be attributed to peripherally derived irisin. tate after other protocols (e.g. MICT, VICT), therefore, the
In support of our findings that the higher-intensity exercise, correlational analyses presented for BDNF and lactate must
SIT, elicited the highest BDNF response, our results further be interpreted with the possibility that the peak blood lactate
show a significant correlation of moderate strength between levels may have been missed.

SPRINT INTERVAL TRAINING INCREASES BDNF Medicine & Science in Sports & Exercise® 1215

Copyright © 2020 by the American College of Sports Medicine. Unauthorized reproduction of this article is prohibited.
 Studies investigating basal BDNF levels in circulation after lifestyle can play a crucial factor in the development of AD
a long-term exercise-training program demonstrate increased and normal age-related cognitive decline, the present study
resting levels. Using HIIT over a 3-month period demon- provides rationale for the use of SIT in provoking the most ro-
strated increased resting BDNF in active young healthy adults bust BDNF response compared with CTRL modalities of
(39). Rodent studies have indicated the exercise-induced lesser intensity in the prevention of brain illness throughout
BDNF mediated brain benefits to be transient during acute ex- the aging process.
ercise as increased hippocampal and cortical BDNF content In conclusion, the present study demonstrates changes in
returns to baseline 4 h after exercise cessation (12). With the plasma BDNF response after acute running-based exercise to
Downloaded from http://journals.lww.com/acsm-msse by BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsIHo4XMi0hCyw

scope of identifying the most robust form of acute exercise be dependent on exercise intensity. Our results demonstrate
to integrate into long-term exercise training and optimize sprint-interval training to be a superior exercise modality in in-
BDNF levels in circulation, our results show SIT to be of most creasing BDNF levels in circulation compared with prolonged
CX1AWnYQp/IlQrHD3i3D0OdRyi7TvSFl4Cf3VC1y0abggQZXdgGj2MwlZLeI= on 05/26/2023

use. This is important considering HIIT and SIT have recently continuous exercise of lower intensity. Moreover, the present
become popular exercise modalities in both recreation and study provides valuable information in targeting exercise mo-
clinical settings and routinely demonstrate benefits to body dalities that elicit a greater blood lactate response in the design
composition and markers of cardiorespiratory health and per- of exercise prescription optimizing for brain health through
formance such as peak oxygen consumption (V̇O2peak) and an- provoking a neuroprotective BDNF response.
aerobic tolerance (39). Moreover, higher-intensity exercise of
shorter durations has shown to be an effective intervention in This work was supported by a Natural Sciences and an Engineer-
symptom management and improving outcomes in patients ing Research Council Discovery Grant to R. E. K. M. (NSERC; grant
RGPIN-2017-03904) and T. J. H. (NSERC; grant RGPIN-2016-06118).
with mental disorders like depression and schizophrenia G. H. is funded by a Master’s NSERC Graduate Scholarship and L. K. T.
(40,41). The overarching purpose of the present study in was supported by an Ontario Graduate Scholarship.
healthy young male adults was to identify the optimal exercise No conflicts of interest to disclose.
The present study do not constitute endorsement by ACSM.
modality in elevating plasma BDNF levels postexercise for The results of the study are presented clearly, honestly, and without
exercise design in obtaining neuroprotective benefit. As fabrication, falsification, or inappropriate data manipulation.

REFERENCES
1. World Health Organization. Towards a Dementia Plan: A WHO 13. Dinoff A, Herrmann N, Swardfager W, et al. The effect of exercise
Guide. 2017. Available from: http://www.who.int/mental_health/ training on resting concentrations of peripheral brain-derived neuro-
neurology/dementia/en/. trophic factor (BDNF): a meta-analysis. PLoS One. 2016;11(9):1–21.
2. Norton S, Matthews FE, Barnes DE, Yaffe K, Brayne C. Potential 14. Rasmussen P, Brassard P, Adser H, et al. Evidence for a release of
for primary prevention of Alzheimer’s disease: an analysis of brain-derived neurotrophic factor from the brain during exercise.
population-based data. Lancet Neurol. 2014;13(8):788–94. Exp Physiol. 2009;94(10):1062–9.
3. Kaplan DR, Miller FD. Neurotrophin signal transduction in the ner- 15. Dinoff A, Herrmann N, Swardfager W, Lanctôt KL. The effect of
vous system. Curr Opin Neurobiol. 2000;10(3):381–91. acute exercise on blood concentrations of brain-derived neurotrophic
4. Vaynman S, Ying Z, Gomez-pinilla F. Hippocampal BDNF mediates factor in healthy adults: a meta-analysis. Eur J Neurosci. 2017;46(1):
the efficacy of exercise on synaptic plasticity and cognition. Eur J 1635–46.
Neurosci. 2004;20:2580–90. 16. Ströhle A, Stoy M, Graetz B, et al. Acute exercise ameliorates re-
5. Manju S, Vignoli B, Canossa M, Blum R. Neurobiology of local and duced brain-derived neurotrophic factor in patients with panic disor-
intercellular BDNF signaling. Pflugers Arch Eur J Physiol. 2017;3: der. Psychoneuroendocrinology. 2010;35(3):364–8.
1–18. 17. Gustafsson G, Lira CM, Johansson J, et al. The acute response of
6. Klein AB, Williamson R, Santini MA, et al. Blood BDNF concentra- plasma brain-derived neurotrophic factor as a result of exercise in ma-
tions reflect brain-tissue BDNF levels across species. Int J jor depressive disorder. Psychiatry Res. 2009;169(3):244–8.
Neuropsychopharmacol. 2011;14(3):347–53. 18. Coelho FG, Gobbi S, Andreatto CAA, Corazza DI, Pedroso RV,
7. Pan W, Banks WA, Fasold MB, Bluth J, Kastin AJ. Transport of Santos-Galduróz RF. Physical exercise modulates peripheral levels
brain-derived neurotrophic factor across the blood–brain barrier. of brain-derived neurotrophic factor (BDNF): a systematic review
Neuropharmacology. 1998;37(12):1553–61. of experimental studies in the elderly. Arch Gerontol Geriatr. 2013;
8. Lommatzsch M, Zingler D, Schuhbaeck K, et al. The impact of age, 56(1):10–5.
APPLIED SCIENCES

weight and gender on BDNF levels in human platelets and plasma. 19. Szuhanya K, Bugattia M, Otto MW. A meta-analytic review of the ef-
Neurobiol Aging. 2005;26(1):115–23. fects of exercise on brain-derived neurotrophic factor. J Psychiatr
9. Krabbe KS, Nielsen AR, Krogh-Madsen R, et al. Brain-derived neu- Res. 2015;60:56–64.
rotrophic factor (BDNF) and type 2 diabetes. Diabetologia. 2007; 20. Ferris LT, Williams JS, Shen CL. The effect of acute exercise on se-
50(2):431–8. rum brain-derived neurotrophic factor levels and cognitive function.
10. Lee BH, Kim H, Park SH, Kim YK. Decreased plasma BDNF Med Sci Sports Exerc. 2007;39(4):728–34.
level in depressive patients. J Affect Disord. 2007;101(1–3): 21. Saucedo Marquez CM, Vanaudenaerde B, Troosters T, Wenderoth
239–44. N. High-intensity interval training evokes larger serum BDNF levels
11. Laske C, Stellos K, Hoffmann N, et al. Higher BDNF serum levels compared with intense continuous exercise. J Appl Physiol. 2015;
predict slower cognitive decline in Alzheimer’s disease patients. Int 119(12):1363–73.
J Neuropsychopharmacol. 2011;14(3):399–404. 22. Cabral-Santos C, Castrillón CIM, Miranda RAT, et al. Corrigendum:
12. Seifert T, Brassard P, Wissenberg M, et al. Endurance training en- “Inflammatory cytokines and BDNF response to high-intensity in-
hances BDNF release from the human brain. AJP-Regul Inter Comp termittent exercise: Effect the exercise volume.” Front Physiol.
Physiol. 2010;298(2):372–7. 2016;7:662.

1216 Official Journal of the American College of Sports Medicine http://www.acsm-msse.org

Copyright © 2020 by the American College of Sports Medicine. Unauthorized reproduction of this article is prohibited.
 23. Gilder M, Ramsbottom R, Currie J, Sheridan B, Nevill AM. Effect of 33. Bergersen LH. Lactate transport and signaling in the brain: potential
fat free mass on serum and plasma BDNF concentrations during ex- therapeutic targets and roles in body-brain interaction. J Cereb Blood
ercise and recovery in healthy young men. Neurosci Lett. 2014;560: Flow Metab. 2015;35(2):176–85.
137–41. 34. Albrecht E, Norheim F, Thiede B, et al. Irisin—a myth rather than an
24. Wrann CD. FNDC5/Irisin—their role in the nervous system and as a exercise-inducible myokine. Sci Rep. 2015;5:1–10.
mediator for beneficial effects of exercise on the brain. Brain Plast. 35. Sanchis-Gomar F, Alis R, Pareja-Galeano H, Romagnoli M, Perez-
2015;1:55–61. Quilis C. Inconsistency in circulating irisin levels: what is really hap-
25. Islam H, Townsend LK, McKie GL, et al. Potential involvement of pening? Horm Metab Res. 2014;46(8):591–6.
lactate and interleukin-6 in the appetite-regulatory hormonal response 36. He Z, Yin S, Tian Y, et al. Myokine response to high-intensity interval vs.
to an acute exercise bout. J Appl Physiol (1985). 2017;123(3):614–23. resistance exercise: an individual approach. Front Physiol. 2018;9:1–13.
Downloaded from http://journals.lww.com/acsm-msse by BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsIHo4XMi0hCyw

26. Gass GC, Rogers S, Mitchell R. Blood lactate concentration follow- 37. Kabasakalis A, Nikolaidis S, Tsalis G, Christoulas K, Mougios V. Ef-
ing maximum exercise in trained subjects. Br J Sports Med. 1981; fects of sprint interval exercise dose and sex on circulating irisin and
15(3):172–6. redox status markers in adolescent swimmers. J Sports Sci.
CX1AWnYQp/IlQrHD3i3D0OdRyi7TvSFl4Cf3VC1y0abggQZXdgGj2MwlZLeI= on 05/26/2023

27. Boström P, Wu J, Jedrychowski MP, et al. A PGC1-α-dependent 2019;37(7):827–32.

myokine that drives brown-fat-like development of white fat and 38. Huh JY, Panagiotou G, Mougios V, et al. FNDC5 and irisin in humans:
thermogenesis. Nature. 2012;481(7382):463–8. I. Predictors of circulating concentrations in serum and plasma and II.
28. Fox J, Rioux BV, Goulet EDB, et al. Effect of an acute exercise bout mRNA expression and circulating concentrations in response to weight
on immediate post-exercise irisin concentration in adults: a meta- loss and exercise. Metabolism. 2012;61(12):1725–38.
analysis. Scand J Med Sci Sports. 2018;28(1):16–28. 39. Murawska-Cialowicz E, Wojna J, Zuwala-Jagiello J. Crossfit training
29. Norheim F, Langleite TM, Hjorth M, et al. The effects of acute and changes brain-derived neurotrophic factor and irisin levels at rest,
chronic exercise on PGC-1α, irisin and browning of subcutaneous after Wingate and progressive tests, and improves aerobic capacity
adipose tissue in humans. FEBS J. 2014;281(3):739–49. and body composition of young physically active men and women.
30. Huh JY, Siopi A, Mougios V, Park KH, Mantzoros CS. Irisin in re- J Physiol Pharmacol. 2015;811–21.
sponse to exercise in humans with and without metabolic syndrome. 40. Dauwan M, Begemann MJH, Heringa SM, Sommer IE. Exercise im-
J Clin Endocrinol Metab. 2015;100(3):E453–7. proves clinical symptoms, quality of life, global functioning, and de-
31. Jabre V, Stephan JS, Nasrallah P, et al. Lactate mediates the effects of pression in schizophrenia: a systematic review and meta-analysis.
exercise on learning and memory through SIRT1-dependent activa- Schizophr Bull. 2016;42(3):588–99.
tion of hippocampal brain-derived neurotrophic factor (BDNF). J 41. Gourgouvelis J, Yielder P, Clarke ST, Behbahani H, Murphy BA.
Neurosci. 2019;1661–18. Exercise leads to better clinical outcomes in those receiving medica-
32. Wrann CD, White JP, Salogiannnis J, et al. Pathway. 2014;18(5): tion plus cognitive behavioral therapy for major depressive disorder.
617–32. Front Psych. 2018;9:37.

APPLIED SCIENCES

SPRINT INTERVAL TRAINING INCREASES BDNF Medicine & Science in Sports & Exercise® 1217

Copyright © 2020 by the American College of Sports Medicine. Unauthorized reproduction of this article is prohibited.