The influence of body posture and added mass on intrinsic and extrinsic foot muscle activation and force output during common foot strengthening exercises

Highlights

• •Across 16 common foot strengthening exercises, muscle activation and metatarsophalangeal joint torque varied substantially.
• •Intrinsic and extrinsic foot muscle activation and metatarsophalangeal joint torque significantly increased with the addition of a forward lean and added mass.
• •Adding a forward lean to exercises produced a 35% median increase in metatarsophalangeal joint torque, exceeding the effect of additional mass.
• •Traditional toe strengthening exercises generated high muscle activation but low torque output, questioning their strength-training effectiveness. Understanding the strengthening capabilities of toe strengthening exercises may be better established focusing on torque production than muscle activation.

Abstract

Background

Intrinsic and extrinsic foot muscles contribute to foot and lower limb function. This knowledge provides opportunities to target these muscles in exercise and rehabilitation. However, a barrier is the limited understanding of how prescribed exercises affect muscle activation and force output. This study examined muscle activation and torque production about the metatarsophalangeal joints of intrinsic and extrinsic foot muscles during common foot strengthening exercises. The secondary aim was to determine whether the addition of a postural change (leaning forward) or addition of mass can further increase muscle activation and torque.

Methods

Fifteen healthy participants (9 males, 6 females; age = 36.5 ± 8.0 years, mean ± SD) participated. Muscle activation was measured using fine-wire electromyography (EMG) electrodes for 2 intrinsic and 2 extrinsic muscles, and surface EMG electrodes for 4 superficial muscles. Metatarsophalangeal joint torque was measured using a custom toe dynamometer. Participants attended one 90-min session and performed 16 exercises. Five included the addition of leaning forward and 3 the addition of mass (20% of participant’s mass) to compare their effects on muscle activity and torque.

Results

Muscle activation and metatarsophalangeal joint torque varied considerably across exercises. Both the addition of leaning forward and the addition of mass increased muscle activation and metatarsophalangeal joint torque, with the addition of leaning forward producing a 35% median increase in torque about the metatarsophalangeal joints.

Conclusion

Many common exercises, despite high muscle activation, produce relatively low metatarsophalangeal joint torque, raising questions about their clinical value. Adding a forward lean significantly increases toe flexor muscle activity and torque, reaching levels greater than added mass and comparable to walking.

Graphical absctract

1. Introduction

The intrinsic and extrinsic foot muscles make significant contributions to the function of the foot. During ambulation these muscles absorb energy during initial foot contact with the ground, generate power for forward propulsion, provide dynamic support for the longitudinal arch, and assist with upright balance control.1, 2, 3, 4 Weakness of these muscles has been associated with numerous foot and lower limb musculoskeletal pain conditions5, 6, 7 and an increased risk of falls in older people,⁸ which highlights their importance.

Strengthening exercises have been found to be effective in reducing pain and improving function for many lower limb musculoskeletal conditions.9, 10, 11, 12 However, unlike musculoskeletal conditions at the knee and hip, evidence to guide exercise-based interventions for musculoskeletal foot pain is sparse.12, 13, 14 As an example, a recent Delphi study conducted by our research group established expert consensus for a set of exercises that were deemed appropriate for the management of plantar heel pain, most of which targeted toe flexion.¹⁵ However, the effectiveness of these exercises to achieve functionally worthwhile increases in strength and alterations in function remains unknown.

Previous studies investigating how different foot strengthening exercises target individual foot muscles have relied on muscle activation measures such as electromyography (EMG) to infer muscle force production.16, 17, 18, 19 A significant limitation of these studies is that EMG does not always provide an accurate estimate of muscle force.^20,21 The capacity of a muscle to produce force for a given level of neural drive (activation) is influenced by the length and lengthening velocity of the muscle as it contracts.^20,21 Thus, a muscle’s force output can drastically change across a range of different muscle lengths, whilst maintaining a similar level of activation. The mismatch between muscle force output and muscle activation is particularly relevant for the toe flexor muscles, due to wide variation in foot position (and underlying muscle lengths) during common foot strengthening exercises (e.g., short foot, toe flexion, and heel raises). A major barrier to developing effective strength-based exercise interventions for the foot is understanding which tasks produce muscle forces that are functionally relevant.

During gait, the intrinsic and extrinsic foot muscles actively lengthen and shorten during each step, optimizing force output from these relatively small muscles.^4,22, 23, 24, 25 This function is achieved via the action of these muscles across the metatarsophalangeal (MTP) joints and the longitudinal arch.⁴ During walking and running, peak intrinsic muscle tendon unit lengths and peak internal foot loading occur in synchrony, corresponding to late midstance when the heel begins to rise from the ground.^1,4,23 Exercise interventions that facilitate force production at longer muscle lengths may be essential to ensure sufficient force is generated to replicate the force requirements of gait.² Common foot strengthening exercises are performed in positions that require muscle force generation at lengths considerably shorter than the operating lengths of these muscles in gait.^18,19,26,27 Therefore, the capacity for the intrinsic and extrinsic foot muscles to develop functionally relevant strength gains from traditional foot strengthening exercises is questionable.

A potential solution to increase force output from the intrinsic and extrinsic toe flexor muscles during strength tasks is to replicate body posture during the period of peak muscle length and force length in walking and running. For example, a simple cue to lean forward shifts the body’s center of mass anteriorly, increasing load through the foot, which flattens the longitudinal arch and stretches the toe flexor muscles (Supplementary Fig. 1).^23,28 This shift in body posture also produces involuntary activation of the intrinsic foot muscles.²⁹ This simple postural adjustment could potentially increase muscle force output, enhancing the functional relevance and effectiveness of many common foot strengthening exercises.³⁰

Furthermore, the prescription of strengthening exercises at the foot and ankle is often performed at low to very low loads,^27,31, 32, 33, 34 which also does not reflect the loading demands placed on the foot during walking or running. During these activities, the loads on the foot frequently exceed body weight,^2,28 underscoring the importance of exercises that replicate these conditions. Therefore, modifying strengthening exercises with increased mass may also provide a solution to achieve functionally relevant loads through the foot during foot strengthening exercises.

Therefore, this study had 4 aims, which were to determine: (a) how a series of commonly prescribed foot and ankle strengthening exercises influence the magnitude of muscle activation and force output in healthy adults, which was the primary aim; (b) if adjusting body posture (by leaning forward) increases toe flexor muscle activation and force output; (c) if adding mass increases toe flexor muscle activation and force output, and (d) if adding a forward lean or adding of mass produce greater increases in muscle activation and force output when performing common weightbearing foot strengthening exercises.

2. Materials and methods

2.1. Participants

Fifteen healthy participants were recruited into this study. Participants were eligible if they had no history of lower limb injury or surgery in the previous 6 months, or known systemic (e.g., diabetes, inflammatory arthritis) or neurological impairments. An a priori calculation was made to determine the number of participants to provide sufficient power (>90%) to detect an effect size of 0.8 (large). Ethics approval was obtained from the University of Queensland Ethics Committee and La Trobe University Human Research Ethics Committee (2023/HE000461). All participants provided written informed consent prior to participation.

2.2. Data collection

2.2.1. Muscle activation

Muscle activation was measured using fine-wire EMG electrodes for the intrinsic toe flexor muscles abductor hallucis (AH) and flexor digitorum brevis (FDB) and the extrinsic toe flexor muscles flexor hallucis longus (FHL) and flexor digitorum longus (FDL). Surface EMG electrodes were applied to record activation patterns of the superficial intrinsic muscle flexor hallucis brevis (FHB), as well as the superficial posterior leg muscles gastrocnemius lateralis (GL), gastrocnemius medialis (GM), and soleus (SOL).

Insertion of the fine-wire electrodes was performed in the right foot and leg of each participant under ultrasound guidance (5–15 MHz, high-frequency linear array, L15 HD₃; Clarius, Vancouver, Canada) in accordance with our previously described approaches (Supplementary Table 1).^22,28 Briefly, fine-wire electrode positions were determined for each intrinsic (AH and FDB) and extrinsic foot muscles (FHL and FDL) by identifying the muscle of interest with ultrasound imaging (5–15 MHz, high-frequency linear array, L15 HD₃; Clarius), informed by the anatomical guides provided in Perotto et al.³⁵ Subsequently, the custom made paired fine-wire electrodes (4 mm recording surface, 50.0 mm × 0.5 needle) were inserted into the muscle of interest under real-time ultrasound guidance. Sterile techniques were used for the insertion of all wires.

EMG electrode positions for the FHB, GM, GL, and SOL were established based on Surface ElectroMyoGraphy for Non-Invasive Assessment of Muscles (SENIAM) guidelines.³⁶ Prior to placement of surface EMG electrodes, the skin was lightly abraded and cleaned with isopropyl alcohol. Surface Ag–AgCl electrodes (Arbo H93SG, Tyco Healthcare Group, Neustadt, Germany) with a diameter of 10 mm and an interelectrode distance of 20 mm were placed on the GM, GL, and SOL muscles based on SENIAM guidelines. The location of the FHB surface electrodes was established based Perotto et al.³⁵ under ultrasound guidance in accordance with our previously described approach.³⁶

Following insertion and placement of EMG electrodes, participants were asked to perform known movements to activate each muscle, to ensure signal quality and correct electrode position. For the fine-wire electrodes, if the EMG signal to noise ratio was low, we slightly adjusted (by ∼1 mm) the position of the fine-wire electrode to achieve improved signal quality. In 1 case, where a participant reported discomfort in the AH, the electrodes were removed and no measurement was recorded from the AH for this participant. In two cases, the signal quality of the surface electrodes for FHB was poor and could not be improved. No FHB EMG measurements were included for these participants.

2.2.2. MTP joint torque

Torque produced about the MTP joints was measured using a previously described, custom-built toe dynamometer (Toe Extension Rotation Device,³⁷ Supplementary Fig. 2). The MTP joint torque was collected as a surrogate external measure of intrinsic and extrinsic toe flexor muscle force, as these muscles all produce a flexion torque about the MTP joints. The MTP joints were positioned in a neutral position (∼0 degrees of extension) with the participant standing on the Toe Extension Rotation Device (Supplementary Fig. 2). As the participant applied a flexion force with the toes, torque measuring strain gauges (187UV; MicroMeasurements, Wendell, NC, USA) within the Toe Extension Rotation Device quantified the ensuing MTP joint torque that was produced.

2.3. Protocol

Participants attended the laboratory for one 90-min session. Prior to data collection, participants were familiarized with the experimental set-up and measurement devices. They received verbal instruction and demonstration of all exercises and had the opportunity to have 3 familiarization trials without electrodes or force plates.

2.3.1. Standard exercise protocol

The exercises were selected based on consensus obtained from a panel of experts, as a strengthening protocol for plantar heel pain (Table 1).¹⁵ Participants were asked to perform the specified exercises in series, completing 3 repetitions of 3 s during each trial. Each repetition was monitored for time with a metronome set at 60 beats per minute. Participants were instructed to perform the contraction on 1 beat, hold for the next 3 beats and release on the 5th beat. Participants rested for 10 s between each repetition. The exercises and their verbal instructions are outlined in Table 1.

Table 1.

Exercises used for assessment.

Name	Description	Image
Quiet stance	Participants were asked to stand quietly and avoid movement.
Squat	Participants were asked to keep their feet in contact with the ground, bend their knees and squat to a position where their thighs were approximately parallel to the floor.
Short foot	Participants were asked to shorten their foot (i.e., lift their arch) by actively attempting to bring the ball of their foot toward their heel, while keeping their toes flat on the floor.
Toe spread out	Participants performed the exercise as a sequence of movements; (a) extending all toes, (b) separating the toes as much as possible, (c) flexing the little (5th) toe, (d) flexing big toe (hallux), and (e) flexing the remaining toes. Each movement occurred on a single beat of the metronome and once all digits are flexed held for three beats before releasing.
Hallux flexion	Participants were asked to flex their big toe (hallux) toward the ground with as much force as possible.
Lesser digit flexion	Participants were asked to flex their smaller (2nd to 5th) toes toward the ground with as much force as possible.
Double leg heel raise	Participants were asked to rise directly upwards, lifting heels off the ground, hold and lower to the starting position. This was performed on 2 feet.
Single leg heel raise	Participants were asked to rise directly upwards, lifting their heel off the ground, hold and lower to the starting position. This was performed on 1 foot.

2.3.2. Adding a lean to increase muscle force

To test our hypothesis that adjusting body posture with an anterior shift in the body’s center of mass (i.e., leaning forwards) would increase muscle activation and muscle force (MTP joint torque), we modified 5 weightbearing exercises from the standard protocol (quiet stance, hallux flexion, lesser digit flexion, double leg heel raise, and single leg heel raise) by asking participants to lean forward prior to initiation of the task. Participants were instructed to “lean forward until you feel your heels begin to lift from the ground”, with this position considered the starting position for each exercise. A stabilizing partition was placed approximately 40 cm in front of the participant to improve stability and eliminate the potentially confounding influence of postural sway on muscle activation and MTP torque. Participants were instructed to only apply light contact force on the partition, with 2 fingers. The degree of force applied was monitored by investigators via visual inspection throughout the execution of the exercise. If the participant was deemed to apply too much force, the attempt was ceased and re-started.

2.3.3. Adding mass to increase muscle force

To test our hypothesis that adding mass could also be an effective means to increase muscle activation and MTP joint torque, we modified three of the previously described exercises (leaning forwards (lean), double leg heel raise, and single leg heel raise). The increase in mass was set to the kettlebell mass that was closest to 20% of each participant’s body mass. As kettlebells come in specified weights of 10, 12, and 16 kg, if a participant weighed 50 kg, for example, they were provided a kettlebell of 10 kg. Participants were instructed to hold the kettlebell comfortably in front of, and maintain contact with, the body.

2.4. Data processing and analysis

2.4.1. Muscle activation

EMG data were sampled at 4000 Hz, amplified 1000 times and band-pass filtered at 30–1000 Hz for fine-wire recordings and 30–500 Hz for surface recordings (MA300; Motion Lab Systems, Baton Rouge, LA, USA). Data were digitally converted with a 14-bit analogue-to-digital converter (Micro3-401; Cambridge Electronic Design, Oxford, UK). Subsequently, EMG signals were high-pass filtered using a 4th-order Butterworth filter at 35 Hz, and root mean square (RMS) EMG signal amplitude was calculated (Spike2; Cambridge Electronic Design, Oxford, UK). For each participant, EMG data were subsequently normalized to the maximal RMS amplitude recorded for each individual muscle, across all trials. Given that numerous maximal effort contractions were performed as part of the protocol, we assumed the maximal signal amplitude to be the maximal voluntary activation, regardless of the task it was obtained from. An average of the RMS amplitude across the plateau of the exercise (3 s hold phase) was calculated for each trial. These 3 values were then averaged again, to provide an average activation measure for each muscle during each exercise task.

2.4.2. MTP joint torque

Torque data were sampled at 4000 Hz, amplified 1000 times using a custom-built amplifier and subsequently converted from analogue-to-digital (Micro3-1401; Cambridge Electronic Design). Similar to the EMG data, torque measurements were normalized to the maximal torque value obtained for each participant, across all exercise trials. An average of the torque amplitude across the plateau of each exercise (3 s hold phase) was calculated for each trial. These 3 values were then averaged again to provide an average MTP joint torque measure for each muscle during each exercise task.

2.5. Statistical analysis

All statistical analyses were conducted using SPSS 28.0 (IBM Corp., Armonk, NY, USA). Normality of data was assessed using a combination of graphical outputs (histogram, box plots, P–P plots and Q–Q plots) and statistical tests (Shapiro-Wilk test, Kolmogorov-Smirnov test, skewness, and kurtosis).

All EMG data were normally distributed. We used a repeated measures one-way analysis of variance to determine the effect of exercise (Aim 1), addition of lean (Aim 2), addition of mass (Aim 3), and the difference between addition of lean and addition of mass (Aim 4) on intrinsic (AH, FHB, and FDB) and extrinsic (FHL, FDL, GL, GM, and SOL) foot muscle activation. After comparison of main effects, we used paired t tests for a post hoc analysis, per muscle, between exercises.

The MTP joint torque data were not normally distributed. We used a Friedman test to determine the overall effect, median differences, interquartile ranges (IQR), and their associated p values to determine the effect of exercise (Aim 1), addition of lean (Aim 2), addition of mass (Aim 3) and the difference between addition of lean and addition of mass (Aim 4) on MTP joint torque production. After comparison of main effects, post hoc analysis between exercises was conducted using a Wilcoxon signed-rank test.

For all analyses the independent variable was the exercise and the dependent variable was the normalized mean EMG muscle activation or median torque production for each exercise across all participants. To account for multiple comparisons across muscles, p values were corrected using the Bonferroni method. A flow chart of the process is provided in Supplementary Fig. 3 and a graphic to support the flow chart is provided in Supplementary Fig. 4.

3. Results

3.1. Participant characteristics

The participants in this study included 9 males and 6 females, age = 36.5 ± 8.0 years, height = 175 ± 12 cm, weight = 72  ± 17 kg; mean ± SD.

3.2. Effect of exercise

EMG and torque data were obtained for all 15 participants across 12 of 16 exercises. Due to issues with data acquisition, data were obtained for 14 of 15 participants for hallux flexion and 13 of 15 participants for the added forward lean and added mass conditions.

3.3. Intrinsic foot muscles (AH, FHB, and FDB) EMG

3.3.1. Effect of exercise on intrinsic muscle activation

There was a main effect of exercise on muscle activation for all intrinsic toe flexors (p < 0.01) (Fig. 1). The exercises with the highest mean EMG RMS amplitude were: single leg heel raise for AH (57.6% ± 19.7%), hallux flexion for FHB (62.5% ± 24.7%), and the single leg heel raise for FDB (58.1% ± 20.6%).

Fig. 1.

Percentage of maximum muscle activation of the intrinsic muscles: (A) abductor hallucis, (B) flexor hallucis brevis, and (C) flexor digitorum brevis for each individual exercise. Exercises of moderate shade indicate modified conditions of added lean and the darkest shade indicate added mass. Box plots represent median, IQR, and mean (white circle). IQR = interquartile ranges.

3.3.2. Effect of adding a forward lean on intrinsic muscle activation

There was a significant main effect of leaning forward on muscle activation during exercise execution (increase with the addition of a forward lean by 11.3% ± 28.0%, p < 0.01) (Fig. 2A). Post hoc comparisons showed activation increased significantly for two muscles with the addition of a forward lean (AH = 11.5% ± 22.9%, and FDB = 14.4% ± 26.2%, both muscles p ≤ 0.01) (Fig.1).

Fig. 2.

Comparison of muscle activation and metatarsophalangeal joint torque production between standing, addition of lean, and addition of mass to exercises for (A) intrinsic muscles, (B) extrinsic muscles, and (C) torque production. Values displayed show percentage of maximum activation or torque output. Box plots represent median, IQR, and mean (white circle). * p < 0.01. IQR = interquartile ranges.

3.3.3. Effect of adding mass on intrinsic muscle activation

There was a significant main effect of added mass on muscle activation during exercise execution for the intrinsic toe flexors (15.9% ± 28.7%, p < 0.01) (Fig. 2A). Post hoc comparisons showed activation increased significantly for one muscle with the addition of mass (AH = 17.6% ± 25.8%, p = 0.015) (Fig. 1).

3.3.4. Comparing the effect of adding a forward lean and adding mass on intrinsic muscle activation

There was no statistically significant difference between the addition of a forward lean and the addition of mass on intrinsic toe flexor muscle activation (p > 0.6) (Fig. 2A).

3.4. Extrinsic foot muscles (FHL, FDL, GL, GM, and SOL) EMG

3.4.1. Effect of exercise on extrinsic foot muscle activation

There was a significant main effect of exercise on muscle activation for the extrinsic toe flexors and ankle plantar flexors (p < 0.01) (Fig. 3). The single leg heel raise produced the highest activation magnitude for all muscles (FHL = 64.0% ± 25.7%, FDL = 70.6% ± 20.1%, GL = 73.4% ± 13.1%, GM = 76.6% ± 10.6%, and SOL = 72.5% ± 19.9%). By comparison, there was very little activation magnitude for traditional toe flexion exercises (e.g. the short foot (FHL = 15.2% ± 18.3%, FDL = 20.3% ± 21.7%, GL = 5.0% ± 8.1%, GM = 3.4% ± 3.7%, and SOL = 8.8% ± 8.6%)).

Fig. 3.

Percentage of maximum muscle activation of the extrinsic muscles: (A) flexor hallucis longus, (B) flexor digitorum longus, (C) gastrocnemius lateralis, (D) gastrocnemius medialis, and (E) soleus for each individual exercise. Exercises of moderate shade indicate modified conditions of added lean and the darkest shade indicate added mass. Box plots represent median, IQR, and mean (white circle). IQR = interquartile ranges.

3.4.2. Effect adding a forward lean on extrinsic muscle activation

There was a significant main effect on muscle activation with the addition of a forward lean (5.3% ± 2.9%, p < 0.01) (Fig. 2B). However, post hoc comparisons revealed that adding a forward lean to an exercise did not significantly increase the activation of individual muscles.

3.4.3. Effect of adding mass on extrinsic muscle activation

There was a significant main effect on muscle activation with the addition of mass (13.1% ± 1.3%, p < 0.01) (Fig. 2B). However, post hoc comparisons revealed that adding mass did not significantly increase the activation of individual muscles.

3.4.4. Comparing the effect of adding a forward lean and adding mass on extrinsic muscle activation

There was a significant difference in muscle activation between adding a forward lean and adding mass (8.2% ± 1.3%, p < 0.01) (Fig. 2B). However, post hoc comparisons revealed that adding mass did not significantly increase the activation of individual muscles.

3.5. Torque

3.5.1. Effect of exercise on torque

The MTP joint torque data obtained during the lesser digit flexion and squat tasks were not normally distributed. Therefore, non-parametric tests were used for analysis of all MTP joint torque data. There was a significant effect of exercise on MTP joint torque production (Friedman’s test statistic χ²(7) = 83.9, p < 0.01) suggesting that some exercises consistently produce a higher torque than others. The MTP joint torque production across all exercises can be observed in Fig. 4.

Fig. 4.

Comparison of metatarsophalangeal joint torque production between exercises. Exercises of moderate darkness indicate modified conditions of added lean and the darkest indicate added mass. Box plots represent median, IQR, and mean (white circle). IQR = interquartile ranges; MTP = metatarsophalangeal.

The exercises that produced the highest MTP joint torques were the single leg heel raise with added lean (97.7% ± 14.4%), the single leg heel raise with added mass (79.2% ± 17.0%) and hallux flexion with added lean (78.8% ± 15.5%). The tasks producing the lowest MTP joint torques were the squat (1.9% ± 2%), the short foot (14.6% ± 12.3%) and the toe spread out (15.2% ± 16.9%) (Fig. 4).

3.5.2. Effect of adding a lean on MTP joint torque

There was a significant main effect on MTP joint torque with the addition of a forward lean (Friedman’s test statistic χ²(4) = 84.1, p < 0.01). The conditions with the addition of a forward lean produced a 25.0% median increase (IQR: 9.8%–46.2%) in MTP joint torque than the same exercise performed with an upright posture (Fig. 2C).

3.5.3. Effect of adding mass on MTP joint torque

There was a significant main effect on MTP joint torque with the addition of mass (Friedman’s test statistic χ²(2) = 49.1, p < 0.01). The added mass conditions produced a 17.8% median increase (IQR: 2.5%–52.5%) in MTP joint torque than the same exercise performed with an upright posture (Fig. 2C).

3.5.4. The effect of adding a forward lean and adding mass on MTP joint torque

There was a significant difference in MTP joint torque production between tasks with an added forward lean and tasks with added mass (Friedman’s test statistic χ²(7) =17.1, p < 0.01). The MTP joint torque was 13.0% higher (median difference, IQR: –1.1% to 22.2%) in tasks that added a forward lean, compared to tasks that added mass (Fig. 2C).

4. Discussion

The aim of our study was to determine if muscle activation and force production differ across a range of common foot exercises. We also sought to determine if modifying body posture (adding a forward lean) or adding mass could further increase muscle activation and force during common weightbearing foot strengthening exercises. As hypothesized, we found that common exercises, such as the short foot and toe spread out, produce high levels of intrinsic foot muscle activation, but very low levels of joint torque (and therefore muscle force). We report that a simple postural cue to lean forward increases muscle activation and MTP joint torque beyond the values reported during tasks with an upright posture. Furthermore, these increases were equal to, or greater than the increases in activation and force observed when participants added 20% of their body mass. Our findings highlight that many traditional foot strengthening exercises may be ineffective for the functional requirements of the foot.

We combined muscle activation with MTP joint torque data to reveal how common foot strengthening exercises influence the neural drive to, and force output from the intrinsic and extrinsic toe flexor muscles. The addition of MTP joint torque measurements provided novel insights that highlight how common, toe-based foot strengthening exercises produce minimal force output, despite high muscle activation levels.^16,18,19,38 The mismatch between muscle activation and force output during exercises such as the short foot, is likely due to the foot positions in these exercises requiring force to be produced at extremely short muscle lengths, impairing force production.^16,18,19,38 Additionally, many of the toe flexion tasks (e.g., short foot and toe spread out) only recruit the intrinsic foot muscles, with very little activation in the extrinsic toe flexors. Interestingly, despite being considered an important exercise inclusion for foot strength programs,¹⁵ the squat produced low levels of foot muscle activation and MTP joint torque. These findings highlight that many common toe-flexion focused foot strengthening exercises are likely to be ineffective in strengthening the foot to meet the functional requirements of walking and running.

A simple modification in body posture (leaning forward) substantially increased intrinsic and extrinsic toe flexor muscle force production. Leaning forward shifts the body’s center of mass anteriorly, increasing loading through the midfoot and MTP joints,¹ causing flattening of the longitudinal arch and stretching of the toe flexor muscle tendon units. This alteration in posture and foot loading enables muscle force outputs that are relevant for the requirements of locomotion. The MTP joint torques produced during the exercise tasks with the addition of a forward lean were comparable to the requirements at the MTP joints during walking (8 Nm),³⁹ but still less than those required for running (∼25 Nm).⁴⁰ Importantly, the cue to lean forward is simple to administer, easy to understand, and requires no expensive equipment. Combining this simple postural adjustment with weightbearing foot strengthening exercises is justified in future clinical studies to evaluate foot strengthening exercises as a target for intervention in foot and lower limb musculoskeletal pathology.

Adding an external mass also increased muscle force production during weightbearing exercises. However, the effect of added mass was less consistent than the effects observed due to leaning forwards. We asked participants to hold a kettlebell in front of their body, which may have introduced a stability challenge, requiring people to shift their center of mass slightly backwards to maintain balance. Therefore, the increased foot loading effects of the added mass may have been offset by a posterior shift in posture that acted to decrease loading through the foot. We did not measure center of mass fluctuations in our study, so this explanation remains speculative. However, compared to the tasks with an added forward lean, the reduced MTP joint torque during the weighted tasks supports this idea.

In future, adding significant mass (beyond 20% of a person’s weight) to a person’s shoulders or back may overcome this challenge. Indeed, the significant increase in mass from double leg heel raise to single leg heel raise may partly explain the significant increase in torque production on the single stance limb.

Our intrinsic and extrinsic foot muscle activation and MTP joint torque data can potentially be applied to guide the selection of foot strengthening exercises for different clinical populations. For example, post-surgical patients may need to initially focus on exercises that target motor control, rather than high force output. In such cases, the toe spread out and short foot exercises may be appropriate as they have a high cognitive load to perform the task and involve high levels of neural drive (muscle activation) but generate low muscle forces and joint torques. However, as patients regain function, exercises can progress to those that are functional more relevant, so they better reflect the increasing capacity of the foot. Muscle force output can be progressively increased by adding a forward lean or adding mass, which significantly increases muscle force requirements. These findings offer a framework for visualizing and planning a progressive strengthening strategy, as outlined in Fig. 5.

Fig. 5.

EMG activation versus torque production: A theoretical model of exercise progressions. Standing exercises are in blue, exercises with a lean are in green, and exercises with added mass are in orange. To obtain intrinsic (abductor hallucis, flexor hallucis brevis, and flexor digitorum brevis) and extrinsic (flexor hallucis longus, flexor digitorum longus, gastrocnemius lateralis, gastrocnemius medialis, and soleus) muscle activation we used the mean activation of all muscles (x axis) and plotted against the median torque production (y axis) for each exercise. The median torque values (Nm) also allow for the comparison of our data to the metatarsophalangeal joint torque expectations of walking, as reported by Farris et al.³⁹ The magnitude of metatarsophalangeal joint torque during walking is shown as a red vertical dotted line, providing context to the exercises in relation to functional expectations. For visualization only, the plot is divided in 9 parts with the darkest green color indicating highest muscle activation and metatarsophalangeal joint torque production and lightest red indicating lowest muscle activation and metatarsophalangeal joint torque production. EMG = electromyography; Nm = Newton meters.

Our findings need to be considered in light of a number of limitations. Firstly, not every exercise was compared with adding a forward lean and adding mass. We excluded tasks such as the short foot and toe spread out because they are cognitively challenging to perform and adding a forward lean or mass would make them even more complex. Secondly, all tests were performed on healthy participants, so we cannot make claims related to the activity or force production in the presence of foot pathology. Thirdly, we have not stipulated the exact degrees required for the addition of a forward lean. Although a more tightly-controlled approach would potentially be beneficial, we intentionally made this description pragmatic, so the approach could be readily implemented into clinical practice. Fourthly, it is possible that different participants may have applied differing degrees of finger pressure to maintain their balance. While this was monitored by the investigators, this may change the degree of force applied by the toes to maintain the addition of a forward lean posture. Finally, we do not know what outcomes these exercises have when provided as part of a long-term strength training plan (i.e., whether they are effective in improving symptoms or overall function).

5. Conclusion

This study has investigated how common foot exercises influence muscle activation and torque production about the MTP joints. Isolated toe flexion and intrinsic foot muscle exercises (i.e., short foot and toe spread out) produced moderate levels of intrinsic muscle activation, low levels extrinsic muscle activation, and low MTP joint torques. The addition of a forward lean increased muscle activation and MTP joint torque production more consistently than the addition of mass. To understand the strengthening benefits between different exercises, MTP joint torque may be more appropriate outcome to focus on, rather than muscle activation.

Authors’ contributions

JWAO and LAK participated in the study conceptualisation and design, data collection, processing, analysis and interpretation, and manuscript preparation; HBM and KBL participated in the study conceptualisation and design, data analysis and interpretation, and manuscript preparation; GAW and MC participated in data interpretation and manuscript preparation. All authors have read and approved the final version of the manuscript, and agree with the order of presentation of the authors.

Declaration of competing interest

All authors declare that they have no competing interests.

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Acknowledgment

The authors would like to thank all the participants who contributed their involvement to this research.