Secondary Logo

Journal Logo

Effect of Segmental, Localized Lower Limb Cooling on Dynamic Balance


Medicine & Science in Sports & Exercise: January 2015 - Volume 47 - Issue 1 - p 66–73
doi: 10.1249/MSS.0000000000000379

Purpose This study aimed to determine the effect of cooling progressively greater portions of the lower extremities on dynamic balance and neuromuscular activation.

Methods Ten healthy males (22.8 ± 3.4 yr, 76.5 ± 9.1 kg) performed one room air temperature control (22.4°C ± 0.8°C) and three trials of cold water immersion at 12°C (lateral malleolus, ankle; lateral femoral epicondyle, knee; anterior superior iliac spine, hip) for 10 min before performing a unipedal balance test (Star Excursion Balance Test (SEBT)) with their dominant limb. Muscle activation of the vastus lateralis, biceps femoris, tibialis anterior, and lateral gastrocnemius was measured with surface EMG during the SEBT.

Results Core temperature remained euthermic throughout all trials. Gastrocnemius temperature decreased from control (30.4°C ± 0.5°C) with knee (23.7°C ± 1.7°C) and hip immersion (22.4°C ± 1.0°C), whereas vastus lateralis temperature decreased from control (33.7°C ± 1.7°C) with hip immersion (27.3°C ± 2.0°C) (P < 0.01 for all comparisons). Cold water immersion influenced mean anterior and posterior reach distance on the SEBT in a dose-dependent fashion. Compared with those in control, mean anterior and posterior SEBT reach distances were not decreased with ankle (−1.38% and −0.74%, respectively) and knee immersion (−2.48% and −2.74%), whereas hip immersion significantly reduced SEBT by 4.73% and 4.05% (P < 0.05, d = 0.52–0.58). Muscle activation was largely unaffected as the lower extremities were cooled, with only the lateral gastrocnemius during the anterior SEBT approaching a decrease (P = 0.059).

Conclusions Cooling larger portions of the lower extremities progressively affect dynamic balance, and thermal protection strategies should focus on maintaining temperature in the large muscle mass of the thigh.

1Environmental Ergonomics Laboratory, Department of Kinesiology, Brock University, St. Catharines, Ontario, CANADA; and 2Department of Life Sciences, University of Roehampton, London, UNITED KINGDOM

Address for correspondence: Stephen S. Cheung, Ph.D., Environmental Ergonomics Laboratory, Department of Kinesiology, Brock University, 500 Glenridge Avenue, St. Catharines, Ontario L2S-3A1, Canada; E-mail:

Submitted for publication March 2014.

Accepted for publication May 2014.

Balance is the ability to maintain the center of mass relative to one’s base of support, and it is controlled through a combination of visual, vestibular, and somatic senses (40). The farther someone is able to reach, the greater their neuromuscular strength, proprioceptive control, and joint range of motion (26). Any disturbance to these senses may impair balance, and one such disturbance may be the reduction in blood flow to the extremities from cold exposure. This redistribution of blood flow may impair the neuromuscular and somatosensory components that are important for execution of kinesthetic tasks such as balance and strength (1,7,9,15). Nerve conduction velocity slows by 1.5–2 m·s−1 for every 1°C reduction in skin temperature (Tsk) (34), whereas for the same drop in muscle temperature (Tmu), muscle spindle firing rates drop by 1–3 pulses per second (7). Furthermore, exteroceptive mechanoreceptors also display temperature-sensitive firing patterns by exhibiting thermoreceptor-like function when exposed to rapid drops in temperature (15). These alterations to the efferent and afferent flow of information within the peripheral nervous system, in addition to increases in joint and muscle viscosity (30), are possible mechanisms for the impaired ability to maintain balance in the cold, as observed by some investigators.

Despite the cold-induced alterations in efferent and afferent feedback, laboratory data into the effects of cold exposure on balance are equivocal (5,10,22,23,28,30,39). Increases in postural sway length (67%–87%), velocity (63%–71%), area (42%–67%) and forward–backward movement (35%–57%) have been observed during a Romberg balance test during whole-body air exposure to cold (10°C), rather than warm (25°C), conditions (23), whereas body sway velocity also increases during localized cooling when individuals are exposed to ice water below the ankle for 20 min (22). In contrast, localized cooling using a crushed ice blanket covering the entire lower limbs reduced the Tmu of the thigh to 31°C but had no effect on Romberg and modified Tandem balance tests (5). Thermoneutral core temperature was maintained in the mentioned studies where rectal (23,30) or tympanic (5) temperature was recorded; therefore, it is likely that any balance impairments that do exist originate from peripheral physiology. It has been suggested that because dynamic postural assessment is more sensitive at detecting changes in balance because of aging, compared with static assessment, this dynamic type of test may also be more sensitive at measuring changes in balance because of temperature differences (5). Perry et al. (28) reported that the ability to remain upright with external perturbations was markedly reduced with a 15-min ice water immersion of only the plantar aspect of the feet. More recently, Piedrahita et al. (30) observed a decrease of 9%–11% in end point excursion during a limits of stability test for individuals who had undergone cold water immersion (CWI) at 15°C up to the popliteal fossa for 1 h. The CWI reduced Tmu and Tsk but did not affect rectal temperature (Tre) (30). Other thermal research has observed that muscle function seems to be impaired in the cold to a greater extent during faster muscular contraction speeds (18,25) and/or when higher muscular strength (4,35) is required.

Most research studying thermal stress on balance manipulates temperature while the body surface area exposed to the thermal environment remains constant. The purpose of this study was to investigate the effects of cooling different limb segments of the lower body on dynamic balance. This was accomplished by comparing the effect of water immersion at 12°C up to the ankle, knee, or hip before the Star Excursion Balance Test (SEBT) with that of a no-cooling control condition trial. The SEBT was initially established as a test to challenge the dynamic balance of more athletic populations (12), requiring participants to stand on one leg while performing a semisquat to reach with the contralateral leg as far as possible. Thus, it measures dynamic balance with a reliance on a squat type of movement pattern common to many athletic board sports—such as alpine skiing, kiteboarding, and water skiing—where cooling may occur during recovery breaks. It was thought that, as more of the limb joints and muscles were affected by the cold, range of motion and neuromuscular control would degrade, which are both important for SEBT task execution (26). Therefore, it was hypothesized that dynamic balance performance (reach distance) would be progressively impaired with greater amount of lower body cooling.

Back to Top | Article Outline



This study was conducted in accordance with the Canadian Tri-Council Policy for the ethical treatment of human participants as part of the Helsinki Declaration II, with approval from the bioscience research ethics board of Brock University (REB #11-030). Ten healthy, recreationally active adult males (age, 22.8 ± 3.4 yr; body mass, 76.5 ± 9.1 kg; height, 1.76 ± 0.06 m; body fat, 12.8% ± 7.7% (mean ± SD)) participated after being thoroughly briefed on the protocol and providing a written informed consent. “Recreationally active” was defined as participating in moderate aerobic and/or strength training (≥3 d·wk−1) for more than 30 min per exercise session. Anthropometric exclusion criteria consisted of body fatness >25%, shank fat volume >1500 cm3, and thigh fat volume >2500 cm3. This was done to increase the likelihood of adequate and consistent cooling from the CWI protocol by accounting for the thermal resistance of adipose tissue. Participants were also excluded if they presented with any neuromuscular, neurological, and balance-related disorders (vision, vestibular, or somatosensory). During experimental trials, participants were asked to refrain from caffeine consumption on the day of the trial and avoid alcohol and excessive (intense aerobic or strength training) lower body exercise 24 h before each test session.

Back to Top | Article Outline

Familiarization and anthropometric assessment

During a familiarization session, participant height and body mass were collected in addition to body fat percentage calculated through a three-site Jackson and Pollock (19) skinfold measurement. Lower extremity segmental volume was calculated through measurement of height and circumference of lower limb anatomical landmarks (20). Fat and lean volume of the shank and thigh was obtained through skinfold measurements of the anterior thigh and medial calf integrated into the segmental lower limb volume measurement. Specifically, the bifold of the skin and subcutaneous fat obtained from the caliper measure (Harpenden; Baty International, West Sussex, United Kingdom) was subtracted from the diameter of the thigh and shank limb segments to generate segmental lean mass diameter. Circumference was then recalculated from this smaller diameter and implemented into the Katch and Katch (20) limb volume equation to obtain a value of lean limb volume. Lean volume was then subtracted from total volume to give fat volume. Leg dominance was determined through asking the participants which leg they would kick a ball with. During the familiarization, participants were also instructed on, and allowed to practice, the balance test protocol.

Back to Top | Article Outline

Experimental protocol

Participants visited the laboratory for four experimental sessions. Trials were conducted with a randomized repeated-measures design and conducted with a minimum of 72-h span between each session to minimize the potential for acclimation to the CWI. While seated within a cooling tank, participants were exposed for 10 min to either a thermoneutral air control or to a 12°C CWI up to the lateral malleolus (ankle), lateral femoral condyle (knee), or anterior superior iliac spine (ASIS) (hip) before the balance test. The water temperature and cooling duration were chosen on the basis of pilot testing with the goal of a significant but moderate drop in local Tmu that would not elicit significant overall core cooling or shivering. The duration was also to simulate recovery time (e.g., chair lifts between alpine runs) in winter or marine balance sports.

All muscle and balance testing occurred within approximately 10 min of completion of the CWI. Upon exiting the immersion tank, participants took a seated position on the edge of a padded bench and the Tmu of the vastus lateralis and gastrocnemius was taken within 2–3 min of the CWI. The dominant leg was secured, and participants then performed maximal isometric voluntary contractions (MVC) for EMG normalization, occurring approximately 3–7 min after CWI. For the SEBT, participants took up position where the floor was taped, with either their big toe or heel adjacent to the start line depending on reach direction for the SEBT. Participants then performed the SEBT barefoot while standing on their dominant limb in anterior and posterior directions. Only these directions were chosen to minimize the overall testing time and potential rewarming of the musculature after CWI. The order of reach direction was kept constant for each participant but was counterbalanced between participants to moderate the temporal effect of rewarming from the CWI and potential local fatigue from the balance task. Local fatigue was also controlled by maintaining consistent rest periods, with 10 s of rest given between each reach attempt and 1 min of rest when switching between reach directions. Timing was accomplished with a stopwatch (Fisher Scientific, Waltham, MA). Participants were free to flex their hip, knee, and ankle to whatever angle they desired. Specifically, their primary instructions were to reach as far as possible without letting their heel come up off the ground and to keep their hands on their hips throughout. During each trial, the two SEBT directions included, at minimum, four practice reaches to minimize any learning effect (32), followed by, at minimum, three test reaches, with the maximum excursion distances recorded through the use of visual spotting and a steel measuring tape. Participants performed more than four practice reaches and more than three test reaches if they lost their balance during any practice or test trial, respectively. The mean number of reach attempts ranged from 7.1 ± 0.3 to 7.7 ± 0.9 and 7.2 ± 0.4 to 8 ± 1.2 for anterior and posterior SEBT directions, respectively, across conditions. The three properly executed test excursion distances for each direction were averaged and then normalized to individual participant’s leg length (ASIS to medial malleolus) (12).

Back to Top | Article Outline

Thermal control

The balance testing was conducted in a laboratory with a mean temperature of 22.0°C ± 0.8°C (relative humidity, 38.8% ± 12.4%). The majority of testing occurred between the months of February and May, with a mean outdoor temperature of 9.5°C ± 6.4°C on trial days. CWI was conducted within a hard plastic tank (1.20 × 0.75 × 0.72 m). Participants maintained a seated position during the cooling, which permitted segmental cooling of the lower extremities (i.e., the thigh remained above the waterline during knee immersion). CWI temperature was held constant at 12°C with a recirculating chiller (model 5202; Polyscience, Niles, IL). During CWI, participants donned the lower half of the outer shell of a waterproof survival suit (Ocean Commander immersion suit; Mustang Survival, Bellingham, WA) to keep both the participant and electronics dry but still exposed to the cold stress. Participants wore a T-shirt and shorts throughout the trial.

Back to Top | Article Outline


Tsk measurements were recorded using wired skin thermistors (NTC 235-1064-ND; Digi-Key, Thief River Falls, MN). These sites included the bulk of the muscle belly of the lateral thigh, lateral shank, and dorsal aspect of the foot. Core temperature was recorded through a Tre probe (Mon-a-therm Core; Mallinckrodt Medical) inserted 15 cm beyond the anal sphincter. The thermistors were tethered to a portable data logger (Smart Reader Plus; ACR Systems, Surrey, Canada). Throughout each trial, the data logger was carried by the participant or placed on the ground while conducting the balance test.

Tmu of the dominant limb vastus lateralis and lateral gastrocnemius was measured with a sterilized 26-gauge Tmu probe (Model MT-26/2; Physitemp Instruments, NJ). Probes were inserted perpendicular to the skin surface into the bulk of the muscle belly after CWI but before balance testing, within 5 cm of the surface EMG (sEMG) electrodes. Tmu was taken at a depth of 3 cm beyond the epimysium on the basis of the point at which resistance from the myofascia was overcome during insertion. Probes were held for approximately 5 s until a stable temperature was recorded and then removed for the MVC and SEBT.

sEMG was recorded on four sites of the dominant limb (vastus lateralis, biceps femoris, tibialis anterior, and lateral gastrocnemius). sEMG recording sites were determined through the location of motor points based on Surface Electromyography for the Non-Invasive Assessment of Muscles recommendations and evoked stimulation. Evoked stimulation was accomplished using stimulating electrodes that were connected in series with an isolation unit (Grass Telefactor SIU8T; Astro-Med, Inc., West Warwick, RI) to a peripheral nerve stimulator (Grass Telefactor S88; Astro-Med, Inc., West Warwick, RI) that delivered a square wave pulse 0.5 ms in duration. Motor points were identified as the locations where the lowest amplitude pulse from the cathode yielded the strongest contraction. Surface electrodes were then placed between this point and the distal tendon of the recorded muscle. These sites were shaved and lightly abraded with isopropyl alcohol and a skin preparation gel (Nuprep®; Weaver and Company, Aurora, CO). Conductive gel (Signagel®; Parker Laboratories Inc., Fairfield, NJ) was then applied liberally to the electrodes before the electrode adhesive backing was removed to ensure that the gel remained localized to the metal contact area, reducing the likelihood of a short cut. The electrodes were then secured to the four sites with Transpore™ tape (3M, St. Paul, MN) in addition to the electrode adhesive. A ground electrode was fixed to the dominant limb ASIS. Between trials, a permanent marker was used to help keep the location for the surface electrodes for all four test conditions.

The study used a hardwired four-channel sEMG system (Delsys Inc., Boston, MA). Muscle activity was acquired through active, single differential electrodes (DE-2.1; Delsys Inc., Boston, MA) with a 10-mm interelectrode distance and a typical common mode rejection ratio of 92 dB transmitted to the Bagnoli amplifier (Delsys Inc., Boston, MA). sEMG activity was recorded within a 250-ms data window that was established around the participant’s maximal reach during each excursion. The root mean square (RMS) amplitude of the sEMG signal during the SEBT was calculated and normalized to the RMS of the EMG signal during an MVC. Analog data were band pass–filtered at 20–450 Hz and then sampled and digitized by a 16-bit A/D DAQ card (National Instruments Corp., Austin, TX). sEMG data were collected through DASYLab (version 10.0; Measurement Computing, Norton, MA), and analysis was performed on the filtered EMG signal using an RMS mathematical protocol through a custom MATLAB® script (Mathworks Inc., Natick, MA). sEMG data were visually inspected before the participant entered the CWI tank to ensure that signal quality was maintained during data collection.

For sEMG normalization, participants sat on the edge of a padded bench with their dominant limb secured with padded straps and their knee flexed at approximately 120° (180°, full knee extension) and a neutral ankle position (0°, minimal plantarflexion and dorsiflexion) with the assistance of a custom-built wooden apparatus. Participants performed an MVC of the vastus lateralis, biceps femoris, tibialis anterior, and lateral gastrocnemius by executing knee extension, knee flexion, ankle dorsiflexion, and ankle plantarflexion, respectively. Participants were verbally encouraged to contract as forcefully as possible for 3 s against the restraining device for each muscle. A 250-ms data window around the largest RMS amplitude value was used for the MVC normalization. Torque values were not measured.

For kinematic analysis (relative joint angle), passive markers were adhered with the 3M Transpore™ tape to eight bony landmarks (dominant limb: head of fifth metatarsal, lateral malleolus, lateral femoral condyle, greater trochanter, ipsilateral acromion process; nondominant limb: head of first metatarsal, medial malleolus, and medial femoral condyle). Two-dimensional kinematic analysis was recorded with a camera (HDR-CX110 Handycam; Sony Electronics Inc., San Diego, CA). The camera was positioned approximately 4.7 m away from the participant and at a height of approximately 0.9 m. Video was recorded at 60 Hz and synchronized to sEMG data using a light-emitting diode/trigger system. As participants performed the balance tasks, the researcher depressed a trigger switch, which activated the light-emitting diode in the field of view of the camera and sent a square wave TTL pulse through the 16-bit multiplex A/D DAQ card (National Instruments Corp., Austin, TX) and ultimately synchronized to the EMG signal within the laboratory computer (DASYLab version 10.0; Measurement Computing, Norton, MA) for analysis.

Back to Top | Article Outline

Data analyses

Mean and SD were used to describe all data. Separate one-factor ANOVA (environmental exposure) with repeated measures was performed for each skin thermistor after CWI during SEBT. Two-factor ANOVA (environmental exposure × SEBT reach direction) and (environmental exposure × muscle site) with repeated measures on SEBT reach distance and Tmu was used to determine whether reach distance was influenced by the amount of peripheral cooling and direction of reach and if significant localized peripheral cooling of muscle was present after CWI. Three-factor repeated-measures ANOVA (environmental exposure × muscle × SEBT reach direction) and (environmental exposure × joint × SEBT reach direction) were used to examine muscle activation and joint angle, respectively, at maximum reach across tests. Bonferroni post hoc comparisons were used to determine where specific differences occurred when a significant main or interaction effect was present. Statistical significance was set at P < 0.05. Cohen d effect sizes were calculated between trial pairings. Cohen’s classification of effect size magnitude was used, whereby d < 0.19 indicates a negligible effect, d = 0.20–0.49 indicates a small effect, d = 0.50–0.79 indicates a moderate effect, and d > 0.8 indicates a large effect. Statistical analysis was performed using the statistical software package SPSS version 16 for Windows (SPSS, Inc., Chicago, IL).

Back to Top | Article Outline


Tre remained stable and euthermic throughout CWI and SEBT testing. All Tsk measurements referenced were taken at the midpoint of SEBT testing. Tsk and Tmu of the immersed lower limb segments were significantly lowered compared with control and when that limb segment was above the waterline during CWI, remaining so throughout SEBT testing (Fig. 1). Thigh Tsk was reduced from a value of 32.3°C ± 0.9°C during the control condition, as well as similar values after ankle and knee immersion, to a temperature of 27.7°C ± 1.7°C (P < 0.01, d = 1.59–1.65 (large effect)) after hip immersion. Tmu of the lateral gastrocnemius was cooled significantly after knee and hip immersion, and vastus lateralis Tmu was also significantly lowered after hip immersion.



Linear kinematic analysis of SEBT reach distance across conditions indicated a significant reduction in SEBT performance with hip immersion when compared with control and ankle immersion (Fig. 2). SEBT reach was reduced on average by 4.73% in the anterior direction and 4.05% in the posterior direction when compared with control (P < 0.05, d = 0.52–0.58 (moderate effect)). When compared with that in the ankle immersion, hip immersion reduced reach distance by 3.40% and 3.34% for anterior and posterior reach directions, respectively (P < 0.05, d = 0.43–0.48 (small-to-moderate effect)). Angular kinematic analysis through motion capture of the ankle, knee, and hip joints did not show any significant change in joint angle at maximum reach of the SEBT in either anterior or posterior directions across environmental conditions (Table 1).





Muscle activation, as measured through sEMG, did not detect any significant changes in muscle activity because of influence from the CWI on SEBT performance from the four muscle groups observed. However, sEMG did detect greater vastus lateralis activity on average relative to the three other muscle groups (biceps femoris, tibialis anterior, and lateral gastrocnemius) during maximal reach in both anterior (P < 0.001) (Fig. 3A) and posterior (P < 0.05) (Fig. 3B) SEBT directions.



Back to Top | Article Outline


This study investigated the influence of segmental cooling on a challenging dynamic balance task and also the potential contributing factors to alteration in task execution. The segmental CWI protocol elicited a gradient of skin and muscle cooling across conditions, with lowered Tsk and Tmu during knee and hip immersion while maintaining a euthermic core temperature throughout all conditions. Hip immersion reduced SEBT reach distance, whereas no significant changes were evident with ankle or knee immersion, and this impairment was observed despite the lack of significant change in joint range of motion and sEMG activity. Thus, our findings add to previous data demonstrating a reduced whole-body balance with leg cooling (30) by finding that cooling of the entire leg is required to significantly affect whole-body dynamic balance.

We used CWI in our protocol to maximize and maintain deep-tissue cooling with each condition, and our calf and thigh Tmu measurements at 3.0-cm depth relative to the epimysium both decreased significantly and also to levels seen in previous reports of neuromuscular impairment with cooling (2). While we measured Tmu before MVC rather than directly before the SEBT, any thermal increase due to the MVC should be constant across conditions. Extensive or intensive cooling of a limb may be required to sufficiently cool muscle to affect balance, and the choice of cooling modality is critical to consider in comparing findings across studies. The magnitude of cooling and also the intensity of the gradient are dependent on the time of exposure and intensity of the thermal stress as well as the amount of thermal resistance (e.g., blood flow, subcutaneous fat) inherent to the individual (2). Greater decrements in nerve motor conduction velocity were reported with 15 min of CWI of the calf compared with that in the application of an ice pack or ice massage (16). This supports previous data demonstrating that 20 min of ice pack application only cooled the upper 1 cm of the vastus lateralis, with minimal and greatly delayed cooling at depths of 2 cm or greater (8). Furthermore, whereas time required to cool the gastrocnemius 1 cm below adipose tissue by 8°C was not different between CWI and a crushed ice bag, intramuscular temperature remained cooler with CWI after 90 min (33).

The effect of local cooling on neuromuscular function and balance may depend on the nature of the task because the body may be able to compensate for changes in individual components involved in muscle function and balance. To extrapolate, contractile characteristics of the first dorsal interosseous muscle were altered with hand CWI, but these changes were not reflected in voluntary submaximal isometric force control (11). In a different study, no change was observed in MVC torque during isometric knee flexion across the thermal conditions when a thermal pack over the anterior thigh was used to manipulate local Tsk from 12.4°C to 40.1°C over 30 min, but tolerance time with a sustained isometric knee flexion was significantly and inversely correlated with thigh Tsk (37). A previous cooling study suggests that faster motor tasks, specifically ones that require greater muscular strength, are more sensitive to temperature (18). Interestingly, larger fast-twitch muscle fibers tend to be found in greater abundance in superficial regions (21) and they are supplied by the largest nerve fiber axons, which also have the greatest temperature sensitivity among neurons (31). Therefore, faster movements and greater force generation that rely heavily on fast-twitch motor units (38) may be more impaired than slower and weaker motor tasks when the body is cooled because of a predisposition to greater temperature reduction in superficial musculature. The SEBT, being a relatively slow dynamic task, may be inherently less sensitive to temperature than tests requiring rapid neuromuscular recruitment, such that only hip CWI cooled the leg musculature and balance systems sufficiently to impair SEBT performance.

Despite our cooling protocol, SEBT was not impaired with ankle or knee CWI and only became significant with full hip CWI. Previous work argues that the SEBT is a motor task that requires a relatively large degree of strength from the quadriceps femoris muscle group (6), although ankle icing has recently been reported to either impair (10) or not influence (39) SEBT performance. Our data indicated that the vastus lateralis muscle was activated on average to 91.1% ± 34.6% of MVC, with the second most active muscle recorded being another thigh muscle group, the biceps femoris at 49.1% ± 38.3% of MVC at maximal reach irrespective of reach direction or environmental condition. Because more strength is required for a motor task higher threshold, larger and more temperature-sensitive motor units are activated (14). Therefore, the greater reliance on fast-twitch motor units within the thigh musculature compared with the shank could explain the significant impairment in SEBT performance with hip but not knee CWI. Furthermore, the effect of CWI on SEBT performance was likely not due to altered microvascularity or reduced perfusion because a 9°C decrease in gastrocnemius Tmu with ice pack cooling did not alter leg blood flow or volume (36).

Although participants anecdotally perceived general increased leg stiffness after CWI, no significant change in active joint range of motion was detected across environmental conditions. As with neuromuscular studies on muscle cooling, flexibility and proprioception data with limb cooling are equivocal. Anterior and posterior cruciate flexibility decreased, and the force required to passively move the knee increased with ice pack cooling compared with that in heating (29). Active range of motion for ankle dorsiflexion decreased after 20 min of 10°C waist immersion (27), whereas 30 min of 14°C waist immersion did not alter knee position sense (3). It is possible that the reduced reach performance that was seen after hip CWI was not due to how far individuals were able to flex at the joints on the stance leg when reaching but may have been due to altered movement patterns within the transverse and frontal planes that were not measured within this study. A previous study analyzing the validity and reliability of recording two-dimensional sagittal joint kinematics from the SEBT with a similar methodology to that of the present study showed reliable (intraclass correlation coefficient 2,1 of 0.76–0.89) measurements of the ankle and knee joints (13). It is possible that individuals were able to compensate for an increase in joint viscosity by increasing triplanar active range of motion at another joint within the kinetic chain. This may be a reason why a similar CWI followed by a limits of stability test where only the ankle was relied on for movement showed significant impairment when the lower extremities were cooled up to the popliteal fossa (30). In contrast, this study showed no significant impairment or reduction in joint range of motion at the same CWI exposure level or greater.

The potential for balance and strength impairment with cold exposure has implications for a variety of populations. For example, aquatic and winter-based board sports, which are increasing in popularity, require high levels of muscular strength in the lower extremities and proficient dynamic balance (17,24). Balance and strength need to remain optimal for performance and for potentially mitigating injury risk, especially in the initial periods of exercise after a prolonged rest (e.g., on chair lifts) before sustained exercise may endogenously increase Tmu. Even when individuals are clothed, it may not be enough of a countermeasure to prevent peripheral cooling. This point was made evident in the current study, as participants wore shorts and waterproof coverings and were still significantly cooled throughout the lower extremities from the CWI. To rule out possible seasonal influences on the observed responses, we subdivided participants into those tested during late winter (February to mid-April: mean outdoor temperature, 5.4°C; n = 5) versus early spring (mid-April to May: mean outdoor temperature, 13.7°C; n = 5). No differences were observed with this analysis, suggesting no seasonal effect from prolonged natural cold acclimatization over the preceding winter. This supports Makinen et al. (23), who reported no difference in postural sway between groups after 10 d of habituation to either 10°C or 25°C air. Further application of the present study may be toward workers and vulnerable or elderly populations exposed to an acute cold stress in winter during everyday activities; however, the data on thermal stress and balance effects on these populations are sparse. It is therefore important that more research regarding the thermal stress on neuromuscular and kinesthetic task performance be carried out. For example, future research may focus on more sedentary or vulnerable populations or else test the effects of fitness and specific balance or exercise training on task performance in extreme environments.

In summary, cooling the lower limb by CWI impairs balance and reduces Tsk and Tmu in a volume-dependent manner. CWI to the hip reduces mean anterior and posterior SEBT reach distances despite the lack of significant change in joint range of motion or sEMG activity; however, CWI to the depth of the ankle or knee does not affect SEBT performance. The current data suggest that the upper leg musculature may be dominant in balance performance with squat-reliant exercise but the mechanisms underlying balance impairment with limb cooling remain unclear.

We express our gratitude to the participants for their efforts throughout the study and to N. Stanov for his enthusiastic assistance with data collection.

The study was supported by the Natural Science and Engineering Research Council of Canada through a discovery grant (#227912-07, S. S. C.). G. L. H. is supported by an Ontario Graduate Scholarship, and S. S. C., by a Canada Research Chair.

The authors have no conflicts of interest to declare. The results of the present study do not constitute endorsement by the American College of Sports Medicine.

Back to Top | Article Outline


1. Asmussen E, Bonde-Peteren F, Jorgensen K. Mechano-elastic properties of human muscles at different temperatures. Acta Physiol Scand. 1976; 96: 83–93.
2. Clarke RS, Hellon RF, Lind AR. The duration of sustained contractions of the human forearm at different muscle temperatures. J Physiol. 1958; 143: 454–73.
3. Costello JT, Donnelly AE. Effects of cold water immersion on knee joint position sense in healthy volunteers. J Sports Sci. 2011; 29 (5): 449–56.
4. Davies CTM, Mecrow IK, White MJ. Contractile properties of the human triceps surae with some observations on the effects of temperature and exercise. Eur J Appl Physiol. 1982; 49: 255–69.
5. Dewhurst S, Riches PE, De Vito G. Moderate alterations in lower limb muscle temperature do not affect postural stability during quiet standing in both young and older women. J Electromyogr Kines. 2007; 17: 292–8.
6. Earl JE, Hertel J. Lower-extremity muscle activation during the star excursion balance tests. J Sport Rehabil. 2001; 10: 93–104.
7. Eldred E, Lindsley DF, Buchwald JS. The effect of cooling on mammalian muscle spindles. Exp Neurol. 1960; 2: 144–57.
8. Enwemeka CS, Allen C, Avila P, Bina J, Konrade J, Munns S. Soft tissue thermodynamics before, during, and after cold pack therapy. Med Sci Sports Exerc. 2002; 34 (1): 45–50.
9. Faulkner JA, Zerba E, Brooks SV. Muscle temperature of mammals: cooling impairs most function properties. Am J Physiol. 1990; 259: R259–65.
10. Fullam K, Caulfield B, Coughlan GF, Delahunt E. The effect of cryotherapy application to the ankle joint on dynamic postural stability in an elite athletic population. Br J Sports Med. 2014; 48 (7): 596.
11. Geurts C, Sleivert GC, Cheung SS. Temperature effects on the contractile characteristics and sub-maximal voluntary isometric force production of the first dorsal interosseous muscle. Eur J Appl Physiol. 2004; 91 (1): 41–5.
12. Gribble PA, Hertel J. Considerations for normalizing measures of the star excursion balance test. Meas Phys Ed Exerc Sci. 2003; 7 (2): 89–100.
13. Gribble PA, Hertel J, Denegar CR, Buckley W. Reliability and validity of a 2-D video digitizing system during a static and a dynamic task. J Sport Rehabil. 2005; 14: 137–49.
14. Henneman E. Relation between size of neurons and their susceptibility to discharge. Science. 1957; 126: 1345–7.
15. Hensel H, Zotterman Y. The response of mechanoreceptors to thermal stimulation. J Physiol (Lond). 1951; 115 (1): 16–24.
16. Herrera E, Sandoval MC, Camargo DM, Salvini TF. Motor and sensory nerve conduction are affected differently by ice pack, ice massage, and cold water immersion. Phys Ther. 2010; 90 (4): 581–91.
17. Hintermeister RA, O’Connor DD, Dillman CJ, Suplizio CL, Lange GW, Steadman JR. Muscle activity in slalom and giant slalom skiing. Med Sci Sports Exerc. 1995; 27 (3): 315–22.
18. Howard RL, Kraemer WJ, Stanley DC, Armstrong LE, Maresh CM. The effects of cold immersion on muscle strength. J Strength Cond Res. 1994; 8 (3): 129–33.
19. Jackson AS, Pollock ML. Generalized equations for predicting body density of men. Br J Nutr. 1978; 40: 497–504.
20. Katch VL, Katch FI. A simple anthropometric method for calculating segmental leg limb volume. Res Q. 1974; 45 (2): 211–4.
21. Knight CA, Kamen G. Superficial motor units are larger than deeper motor units in human vastus lateralis muscle. Muscle Nerve. 2005; 31: 475–80.
22. Magnusson P, Enborn H, Johansson R, Pyykko I. Significance of pressor input from the human feet in anterior-posterior postural control. The effect of hypothermia on vibration-induced body-sway. Acta Otolaryngol. 1990; 110: 182–8.
23. Mäkinen TM, Rintamäki H, Korpelainen JT, et al. Postural sway during single and repeated cold exposures. Aviat Space Envir Med. 2005; 76: 947–53.
24. Mullins NM. Slalom water skiing: physiological considerations and sport conditioning. Strength Cond J. 2007; 29 (4): 42–54.
25. Oksa J, Rintamaki H, Mäkinen T, Hassi J, Rusko H. Cooling-induced changes in muscular performance and EMG activity of agonist and antagonist muscles. Aviat Space Envir Med. 1995; 66: 26–31.
26. Olmsted LC, Carcia CR, Hertel J, Shultz SJ. Efficacy of the star excursion balance tests in detecting reach deficits in subjects with chronic ankle instability. J Athl Train. 2002; 37 (4): 501–6.
27. Patterson SM, Udermann BE, Doberstein ST, Reineke DM. The effects of cold whirlpool on power, speed, agility, and range of motion. J Sports Sci Med. 2008; 7: 387–94.
28. Perry SD, McIlroy WE, Maki BE. The role of plantar cutaneous mechanoreceptors in the control of compensatory stepping reactions evoked by unpredictable, multi-directional perturbations. Brain Res. 2000; 877: 401–6.
29. Petrofsky JS, Laymon M, Lee H. Effect of heat and cold on tendon flexibility and force to flex the human knee. Med Sci Monit. 2013; 19: 661–7.
30. Piedrahita H, Oksa J, Rintamäki H, Malm C. Effect of local leg cooling on upper limb trajectories and muscle function and whole body dynamic balance. Eur J Appl Physiol. 2009; 105: 429–38.
31. Racinais S, Oksa J. Temperature and neuromuscular function. Scand J Med Sci Sports. 2010; 20 (3 Suppl): 1–18.
32. Robinson RH, Gribble PA. Support for a reduction in the number of trials needed for the star excursion balance test. Arch Phys Med Rehab. 2008; 89: 364–70.
33. Rupp KA, Herman DC, Hertel J, Saliba SA. Intramuscular temperature changes during and after 2 different cryotherapy interventions in healthy individuals. J Orthop Sports Phys Ther. 2012; 42 (8): 731–7.
34. Rutkove SB. Effects of temperature on neuromuscular electrophysiology. Muscle Nerve. 2001; 24: 867–82.
35. Sargeant AJ. Effect of muscle temperature on leg extension force and short-term power output in humans. Eur J Appl Physiol Occup Physiol. 1987; 56: 693–8.
36. Selkow NM, Day C, Liu ZQ, Hart JM, Hertel J, Saliba SA. Microvascular perfusion and intramuscular temperature of the calf during cooling. Med Sci Sports Exerc. 2012; 44 (5): 850–6.
37. Thornley LJ, Maxwell NS, Cheung SS. Local tissue temperature effects on peak torque and muscular endurance during isometric knee extension. Eur J Appl Physiol. 2003; 90: 588–94.
38. Wakeling JM, Uehli K, Rozitis AL. Muscle fibre recruitment can respond to the mechanics of the muscle contraction. J R Soc Interface. 2006; 3 (9): 533–44.
39. Williams EE, Miller SJ 3rd, Sebastianelli WJ, Vairo GL. Comparative immediate functional outcomes among cryotherapeutic interventions at the ankle. Int J Sports Phys Ther. 2013; 8 (6): 828–37.
40. Winter DA. Human balance and posture during standing and walking. Gait Posture. 1995; 3: 193–214.


© 2015 American College of Sports Medicine