Intrinsic Foot Muscle Activation During Specific Exercises: A T2 Time Magnetic Resonance Imaging Study

We performed a descriptive laboratory study using a crossover study design. The independent variable was time (pre-exercise and postexercise). Four exercises were analyzed: short-foot exercise, toes spread out, first-toe extension, and second- to fifth-toes extension (Figure 1). Our dependent variable was T2 time, which is presented as the percentage increase in activation (relative to baseline levels of activation) after each of the 4 exercises, recorded individually for these intrinsic plantar foot muscles: abductor hallucis, flexor digitorum brevis, abductor digiti minimi, quadratus plantae, flexor digiti minimi, adductor hallucis oblique, flexor hallucis brevis, and interossei and lumbricals.

Eight National Collegiate Athletic Association Division I cross-country and track athletes (5 men and 3 women: age = 20 ± 0.93 years, height = 180.98 ± 10.84 cm, mass = 70.91 ± 7.82 kg) participated. Participants were included if they had prior experience with the exercises assessed in this study and could perform all 4 exercises correctly on both feet. They were excluded if they had any conditions known to adversely affect muscle size or function or if they were restricted from full participation in sport due to injury status at the time of testing. All participants provided informed consent, and the study methods were approved by the Institutional Review Board for Health Sciences Research.

Participants were scanned on a 3-Tesla MRI machine (Magnetom Trio; Siemens, Munich, Germany) from the most posterior aspect of the calcaneus through the toes with a large Siemens 4-channel flex coil secured around their feet. Images were acquired using a turbo spin-echo pulse sequence with a repetition time of more than 2.5 seconds and echo times of 11 and 67 milliseconds. We performed a voxel-by-voxel fit using these 2 echo times to estimate T2 time. Slice thickness was 10 mm with a 0-mm interslice gap, and scan time was approximately 7 minutes.

Participants sat in a non–weight-bearing position for 10 minutes before the baseline scan to minimize activity of the intrinsic foot muscles. They were transferred from a seated position to the scanner and were positioned supine with their feet first in the scanner. Participants' feet were secured in a standardized position of 0° of ankle dorsiflexion against a footplate and lightly wrapped in a flex coil. The standardized position was selected to ensure consistency across all scans and all participants.

After the baseline scan, participants were returned to a seated position with their feet on the floor. They performed 1 of the 4 exercises with the right foot and a different exercise with the left foot (different pairs of exercises were performed on day 1 and day 2). Exercises were performed in alternating fashion (right foot contraction followed by left foot contraction) to the beat of a metronome while seated. The metronome was set at 20 beats per minute, so that each contraction would be 3 seconds in duration. Participants completed 40 repetitions with each foot and then immediately returned to the scanner for the postexercise scan. We determined the number of repetitions (40) through pilot testing and selected this number to elicit a moderate amount of muscle activation across all 4 exercises in all participants. The order of exercises and the foot (right or left) that performed each exercise were randomized using a Latin square. Participants performed 2 of the exercises (1 with each foot) on day 1 and then returned 48 hours later to complete a new baseline scan, the other 2 exercises, and a postexercise scan.

The short-foot exercise was performed by shortening and raising the medial longitudinal arch by bringing the metatarsal heads toward the calcaneus without flexing the toes or contracting the extrinsic foot muscles.¹ 

The toes-spread-out exercise was performed by extending all 5 toes and then simultaneously abducting all 5 toes while also flexing the first and fifth toes to the ground, keeping toes 2 to 4 extended. The middle toes were then relaxed.¹³ 

First-toe extension was performed by extending the great toe while the second to fifth toes remained on the floor in a neutral position. Second- to fifth-toe extension was performed by extending the second to fifth toes while the great toe remained flat on the floor.²³ 

A specialized software package (version 2.011; WinVessel, Michigan State University, East Lansing, MI) was used to process the MRI scans pre-exercise and postexercise. For each muscle, the series of 3 contiguous MRI slices that provided the largest cross-sectional area was used for analysis. Muscles were identified by a single trained researcher who was blinded to whether scans were taken pre-exercise or postexercise and which exercise was associated with each scan. Baseline scans were used to establish a threshold for muscle activation above resting values. To establish a threshold value for signal intensity, 5 individual circular regions of interest were traced within the desired muscle on each of the 3 baseline slices. The mean T2 value and associated standard deviation were calculated from these 5 regions of interest and then used as the lower threshold. Any pixel that exceeded the lower threshold (T2 mean + 1 SD) was considered active tissue (Figure 2). This threshold was applied to both the pre-exercise and postexercise scans to identify the change in muscle activation after each exercise was performed. An upper limit for signal intensity was set from 900 to 10 000 so that inactive tissue, such as fat, would be excluded from the total number of pixels within each muscle. A lower limit for signal intensity was applied to all scans for any pixel with a value of 0 to 200, which represented other inactive tissues such as blood vessels. Each muscle was manually outlined and a pixel-by-pixel count was obtained for the total number of pixels within that muscle, within the active range (tissue that exceeded 1 SD above the mean resting T2 value to the upper limit of 900), and within any inactive tissue (T2 values between 0 and 200). The percentage increase in muscle activation after each exercise was calculated using the following equations:

This process was followed for each of the 3 largest slices within a muscle, and we used the most active individual slice of each muscle for data analysis. Due to the 10-mm slice thickness and the orientation of the adductor hallucis transverse, we were unable to clearly distinguish this muscle in 3 contiguous slices for each participant and, thus, this muscle was not included in our analysis. In addition, the interossei and lumbricals were not separated by clear fascial borders and, therefore, were treated as a single group in the analysis.

The T2 time for each muscle was measured pre-exercise and postexercise and reported as the percentage increase in muscle activation after each exercise, as demonstrated in the aforementioned equations. Group means and 95% confidence intervals (CIs) were calculated for the percentage increase in activation of each muscle after each exercise. A muscle was considered to be activated above baseline value when the lower limit of the 95% CI did not cross zero. A Pearson correlation coefficient was also calculated to assess possible relationships between pre-exercise levels of activation and the percentage change in activation after the exercises.

Each of the 4 exercises resulted in activation of all of the intrinsic plantar foot muscles by at least 8.9% (range, 8.9% to 35.2%) with associated 95% CIs that did not cross zero (Table).

The short-foot exercise resulted in an increase in muscle activation ranging from 16.7% ± 12.1% in the quadratus plantae to 34.9% ± 8.1% in the abductor digiti minimi (Figure 3A).

Similarly, the toes-spread-out exercise demonstrated an increase ranging from 17.3% ± 10.8% for the interossei and lumbricals to 35.2% ± 11.6% for the abductor digiti minimi (Figure 3B).

The first-toe–extension exercise had a more narrow range for the increase across muscles, with the lowest value being 13.1% ± 8.8% increase for the flexor hallucis brevis and the highest being 18.1% ± 8.6% increase for the flexor digitorum brevis (Figure 3C).

The second- to fifth-toe–extension exercise resulted in an increase that ranged from 8.9% ± 5.0% for the interossei and lumbricals to 22.5% ± 12.6% for the abductor digitorum minimi (Figure 3D).

In addition, heat maps (Figure 4) were created after each exercise depicting the percentage increase in activation for all muscles of each participant. Average percentage increases for each muscle shown on the heat maps were transposed onto corresponding intrinsic foot muscle images (Figure 5) for each exercise. The correlation between pre-exercise activation and the pre-to-post increase in activation was weak (r = −0.19).

All 4 exercises activated all of the intrinsic plantar foot muscles, with average percentage increases in activation that ranged from 8.9% to 35.2%. To our knowledge, this is the first study to describe intrinsic foot-muscle activity using T2 MRI and the first study to quantify the extent of activation of those muscles after specific exercises designed to target them. These findings are important and relevant for clinicians involved in the assessment and treatment of foot and lower extremity conditions that may involve dysfunction of the intrinsic foot muscles and in the prevention of lower extremity injury.

The short-foot exercise showed the greatest mean percentage activation in the abductor digiti minimi (34.9%), abductor hallucis (29.7%), and flexor digitorum (24.8%). This could be a result of the parallel anatomical orientations these muscles have to the longitudinal arches in the foot. The toes-spread-out exercise showed the highest percentage mean activation in the abductor digiti minimi (35.2%), adductor hallucis oblique (31.5%), and flexor digiti minimi (30.2%). Due to the circumferential motion that occurs at the first and fifth toes as these digits are extended, abducted, and then flexed back to the floor, it is not surprising that the intrinsic plantar foot muscles that insert on the first and fifth toes demonstrated the greatest levels of activation with the toes-spread-out exercise. The first-toe–extension exercise showed the greatest percentage mean activation in the flexor digitorum brevis (18.1%), abductor hallucis (16.9%), and flexor digiti minimi (16.3%). As the first toe is extended, the intrinsic plantar flexors for the second to fifth toes are likely being used to keep these toes flat on the floor. Similarly, the abductor hallucis may be activated to either stabilize the great toe during extension or keep the first metatarsal head on the floor. The second- to fifth-toe–extension exercise had the highest mean percentage activation in the abductor digiti minimi (22.5%), flexor hallucis brevis (20.2%), and adductor hallucis oblique (17.6%). For the completion of this exercise, the flexor hallucis brevis and adductor hallucis oblique may play roles in stabilizing the great toe on the ground as the other 4 toes extend.

We did not identify any noticeable trends among participants in terms of which muscles were activated the most or least during each of the 4 exercises. Unfortunately, this limits our ability to make specific recommendations about which exercise may be best suited to target specific intrinsic foot-muscle insufficiencies. However, the distinct variability in the extent of muscle activation across participants underscores several key considerations when interpreting the percentage activation after submaximal exercises. Although we can conclusively identify which muscles were being used for the specific exercises, we cannot speculate about whether increased or decreased levels of activation among participants indicated how difficult the task was for each individual to complete or the muscle force required to complete the exercise. For example, 1 participant may have better neuromuscular control (able to recruit a greater proportion of the motor-neuron pool) of the intrinsic foot muscles and require less activation to complete the 40 repetitions than participants with worse neuromuscular control. Conversely, if a participant has better neuromuscular control, he or she may also be capable of activating more of the motor-neuron pool over the course of 40 repetitions. In addition, the different strategies that could be used to complete these exercises should also be considered when interpreting the present results. For example, the second- to fifth-toe–extension exercise is designed to target the intrinsic plantar foot muscles attached to the hallux. As the participant extends the second through fifth toes off the ground, the muscles acting on the hallux can be activated in an attempt to resist first-toe extension from activation of the extensor hallucis longus, which often occurs simultaneously with activation of the extensor digitorum longus. However, as an individual's fine motor skill improves, he or she could, in theory, accomplish this task by selectively activating the extensor digitorum longus without concurrent extensor hallucis recruitment, which would minimize the activation required of the plantar intrinsic muscles acting on the hallux.

Previous researchers¹³  used sEMG to compare muscle activation of the abductor and adductor hallucis when participants with mild hallux valgus performed the short-foot and toes-spread-out exercises. In that study,¹³  the toes-spread-out exercise showed greater activation of the abductor hallucis than did the short-foot exercise (mean difference = 44.96%). Whereas it was not our aim to compare the extent of activation across exercises, the average percentage increases of the abductor hallucis for the short-foot and toes-spread-out exercises had similar means and associated 95% CIs with substantial overlap, suggesting that the magnitude of difference between these 2 exercises was minimal in our healthy sample. Our results demonstrated that all of the intrinsic plantar foot muscles were activated when both exercises were completed and, thus, crosstalk from deeper or adjacent muscles may have influenced previous sEMG results or differences between healthy and injured populations may need to be examined further.

Kelly et al⁵  recently demonstrated that the extent of activation of the intrinsic foot muscles increases as postural demands increase up to 150% of normal body weight. They also showed⁵  that when electrically inducing a contraction of the muscles of the medial longitudinal arch, arch height and length during weight bearing could be restored to similar dimensions as during non–weight bearing. These findings were instrumental in illustrating the collective role of the intrinsic foot muscles in maintaining the structure of the medial longitudinal arch. Our results have now demonstrated that, regardless of which motion segment of the foot is being moved during the execution of these 4 exercises, all of the intrinsic plantar foot muscles were activated after the task was completed. This finding is important because it indicates how the functional roles of the intrinsic foot muscles cannot simply be described by their anatomical origin and insertion. Our results support the idea that the intrinsic foot muscles act as functional units during exercises² ; however, we cannot identify the exact functional contribution (eg, eccentric control, isometric stability) of each muscle. Although the mechanisms for coordinated motor control among the intrinsic foot muscles are not fully understood, our results suggest that their shared neural innervation may make it more difficult to isolate individual muscles; however, this should be evaluated further.

To our knowledge, we were the first to specifically examine activation of the intrinsic foot muscles through the use of T2 MRI after performing exercises designed to target those muscles. A previous feasibility study²⁴  showed that T2 changes were greater in young participants' foot muscles (105.8% ± 3.3%) than in those of elderly participants (103.9% ± 3.2%). The authors²⁴  also found a longer time to peak T2 signal intensity in elderly participants (157.1 ± 109.9 seconds) than in younger ones (95.1 ± 77.6 seconds), suggesting that age may influence T2 peaks and the time to return to baseline. Heat mapping through T2 percentage-activation change has shown²⁵  varying activation levels in the medial gastrocnemius (13.3% ± 1.3%), lateral gastrocnemius (7.8% ± 1.2%), and soleus (6.3% ± 2.1%) after unilateral heel raises, which are comparable to the percentage increases in the intrinsic plantar foot muscles in our study. Emerging evidence has shown that for T2, higher-intensity exercises result in a quicker rate of change²⁶  as well as a larger overall increase²¹  when compared with lower-intensity exercises. In our study, all exercises were performed without resistance to the beat of a metronome in an attempt to standardize the rate of contraction and contraction intensity across participants. When performing intrinsic foot-muscle exercises, it is common to increase task difficulty by progressing from a seated position to double-leg stance and finally to single-leg stance and, thus, it may be beneficial to examine activation levels of the intrinsic foot muscles throughout this progression. In addition, it would be worthwhile to use T2 MRI to compare the intrinsic foot-muscle function of healthy and injured (plantar fasciopathy, hallux valgus, or other foot and ankle disorders) populations and in tasks such as walking, running, or other functional exercises so that we can expand our evidence-based recommendations for intrinsic foot-muscle exercises.

Our study had some obvious limitations due to the nature of T2 MRI. The overall cost of the MRI scans and the large amount of data-processing time required the relatively small sample size. Our participants were a relatively homogeneous, healthy sample of athletic young adults, and this may limit our ability to generalize our results to less physically active or injured populations as well as to individuals with different arch morphologies.

All of the plantar intrinsic muscles assessed in our study showed increased activation with each of the 4 exercises. The short-foot exercise, toes-spread-out exercise, first-toe–extension exercise, and second- to fifth-toe–extension exercise each caused activation of all of the intrinsic plantar foot muscles and can be recommended as a means of exercising these muscles.