A Normative Dataset of Thumb Muscle Fascicle Lengths in Healthy, Young Adults: An Extended Field-of-View Ultrasound Study

Abstract

Understanding in vivo thumb muscle architecture is essential for advancing musculoskeletal modeling and identifying deviations linked to pathologies. As thumb muscle architecture has primarily been studied in cadavers, the objective of this study was to establish normative data on thumb muscle fascicle lengths in a young, healthy population using extended field-of-view ultrasound (EFOV-US). Six thumb muscles [abductor pollicis brevis (APB) and longus (APL); extensor pollicis brevis (EPB) and longus (EPL); flexor pollicis brevis (FPB) and longus (FPL)] and one wrist extensor [extensor carpi ulnaris (ECU; for comparison purposes only)] were imaged in 18 healthy adults (8 female; age: 22.7 ± 2.0 years; height: 172.1 ± 8.8 cm; weight: 79.0 ± 16.5 kg) [mean ± SD]. Measured fascicles were compared to cadaveric data (all thumb muscles) and ultrasound data (APB, ECU, FPL). Mean fascicle lengths (±SD) were 6.5 ± 0.8 cm (FPL), 3.8 ± 0.4 cm (APL), 4.7 ± 0.5 cm (EPL), 3.7 ± 0.5 cm (EPB), 4.5 ± 0.5 cm (APB), 3.6 ± 0.4 cm (FPB), and 4.2 ± 0.5 cm (ECU). The consistency of our measurements is indicated by the small standard deviations within (±0.1 to ±0.7 cm) and across (±0.4 to ±0.8 cm) participants. Measurement repeatability is high, as demonstrated by low coefficients of variation (range: 0.04–0.08) for the measured thumb muscles. We also examined to what extent anthropometric measurements can be used to predict fascicle lengths and found some significant relationships; however, these relationships were not consistent across all muscles. This study importantly expands our understanding of the complex anatomy of the healthy thumb and provides normative data for future work evaluating hand pathologies.

1. INTRODUCTION

Normative datasets provide essential reference values for understanding muscle structure and function. These datasets enhance the accuracy of musculoskeletal modeling and can aid in the identification of pathologies through deviations from normative values. For example, prior in vivo work collected diffusion tensor imaging data of lower limb muscle architecture and integrated these subject-specific parameters into musculoskeletal models, thereby improving simulation accuracy and individualization of muscle force predictions (Charles et al., 2019, 2020). A study by Handsfield et al. (2014) established normative magnetic resonance imaging (MRI) based volumes for 35 lower limb muscles and developed scaling relationships to body dimensions to facilitate comparisons across individuals of different sizes. This dataset has since been used to identify pathological changes in muscle volumes, such as those seen in individuals pre- and post- anterior cruciate ligament reconstruction (Norte et al., 2018) or in individuals with cerebral palsy (Handsfield et al., 2016).

Despite widespread use of in vivo imaging to study the lower limb, imaging remains underutilized in the upper limb. As a result, cadaveric dissection has long been used to study upper limb muscle structure (An et al., 1981; Brand et al., 1981; Jacobson et al., 1992; Langenderfer et al., 2004; Lieber et al., 1990, 1992). These studies provide insight into fascicle lengths in a small sample. However, they typically provide limited or no demographic data (e.g., age, sex, health history), which may reduce the applicability of findings to the general population. More recently, imaging techniques, such as MRI and ultrasound, have allowed for in vivo examination of muscle structure. Although fascicle length data from upper limb MRI studies remain sparse, MRI has been used to assess other structural parameters including muscle volume (Holzbaur et al., 2007; Saul et al., 2015), muscle thickness (Abe et al., 2018; Juul-Kristensen et al., 2000), and muscle length (Holzbaur et al., 2007; Juul-Kristensen et al., 2000). Ultrasound imaging, which offers a noninvasive, cost-effective method for in vivo fascicle length measurement has become increasingly popular in upper limb studies (Adkins et al., 2017; Adkins & Murray, 2020; Brorsson et al., 2008; Li et al., 2007; Li & Tong, 2005; Nelson et al., 2016, 2018; Rakauskas et al., 2023; Stevens et al., 2014). However, methodological approaches to fascicle length measurement using ultrasound vary across studies. Some employ static B-mode ultrasound imaging to estimate fascicle lengths through linear extrapolation (Brorsson et al., 2008; Li et al., 2007; Li & Tong, 2005; Stevens et al., 2014), while others utilize extended field-of-view ultrasound (EFOV-US) imaging to directly visualize and quantify entire fascicles (Adkins et al., 2017; Adkins & Murray, 2020; Nelson et al., 2016, 2018; Rakauskas et al., 2023). EFOV-US has enhanced measurement accuracy by capturing longer fascicles within a single image and reducing assumptions about fascicle linearity.

Despite the expanding body of literature on upper limb muscle structure, the thumb remains understudied. Most data on thumb muscle fascicle lengths still come from small sets of cadaveric specimens (Amis et al., 1979; Brand et al., 1981; Cutts et al., 1991; Jacobson et al., 1992; Lieber et al., 1992; Ruggiero et al., 2016). This is problematic because reported fascicle lengths for each muscle are widely variable across these studies. For example, fascicle lengths range from 2.5 to 5.5 cm (75% difference) for the extensor pollicis brevis, 3.3 to 5.8 cm (55% difference) for the abductor pollicis longus, and 3.3 to 5.7 cm (53% difference) for the extensor pollicis longus. In this context, in vivo measurements offer a promising approach to overcome these limitations by capturing muscle architecture in its functional state. However, collecting such data is particularly challenging due to the anatomical complexity of the thumb. Nine muscles cross the carpometacarpal joint, including long extrinsic and small intrinsic muscles, which are tightly packed anatomically and difficult to distinguish. Standard ultrasound probes have a field-of-view of 50 mm, yet 90% of upper limb muscles have optimal fascicle lengths that exceed this limit (Adkins & Murray, 2020). EFOV-US enables the measurement of longer muscle fascicles (≥ 50 mm) (Adkins et al., 2017; Adkins & Murray, 2020). However, to date, only one EFOV-US study has examined fascicle lengths in thumb muscles, focusing on just two of the nine muscles (Rakauskas et al., 2023).

Establishing normative data across a larger set of thumb muscles would provide a more comprehensive foundation for understanding muscle structure. Furthermore, as collecting in vivo data is labor-intensive, there is potential to develop regression models based on easily obtainable anthropometric measurements. Rakauskas et al. (2023) reported a significant correlation between fascicle length and hand length. However, their study included only eight participants and examined just two thumb muscles. Expanding the sample size and including additional muscles is necessary to improve generalizability. Such models could complement normative datasets and offer a practical alternative for estimating fascicle lengths in both clinical and research settings.

In this context, the objectives of this study are to (1) establish normative data on thumb muscle fascicle lengths in a young, healthy population using EFOV-US, (2) compare these values with previously reported cadaveric and in vivo measurements, and (3) evaluate the potential to estimate fascicle lengths from anthropometric measurements. To achieve these goals, we imaged seven muscles. Four of these muscles had not been previously studied with EFOV-US, and the other three were selected to enable comparisons with existing in vivo literature. We hypothesized that fascicle lengths measured in this study would be more closely aligned with prior EFOV-US data than with cadaveric measurements. Additionally, we hypothesized that anthropometric measurements would predict fascicle lengths.

2. METHODS

2.1. Participants

Eighteen healthy adults (8 female, 10 male, age: 22.7 ± 2.0 years, height: 172.1 ± 8.8 cm, weight: 79.0 ± 16.5 kg) [mean ± SD] participated in this IRB-approved study (University of Florida, IRB #201802358). All participants reported no arm muscle or joint pain, no medical treatment in the past six months, and no history of neurological or musculoskeletal disorders. All measurements were captured from the participants’ dominant arm. Dominance was determined based on participants’ writing hand. Among participants, one male and one female were left-handed.

2.2. Anthropometric Measurements

Anthropometric data were recorded from each participant and used to assess whether hand and arm dimensions could estimate fascicle lengths. In a manner equivalent to Rakauskas et al. (2023), measurements included forearm length (23.7 ± 1.7 cm; elbow to wrist flexion creases), hand length (18.2 ± 1.3 cm; wrist flexion crease to tip of third finger), hand width (8.6 ± 0.6 cm; widest portion of the hand perpendicular to the forearm, stopping at the metacarpophalangeal joint crease), and thumb length (13.3 ± 0.7 cm; base of thumb where it meets the wrist to the tip).

2.3. Muscle Selection

For each participant, seven muscles [abductor pollicis brevis (APB) and longus (APL), extensor carpi ulnaris (ECU), extensor pollicis brevis (EPB) and longus (EPL), flexor pollicis brevis (FPB) and longus (FPL)] were imaged. The selected muscles included two thumb muscles (APB, FPL) and one wrist extensor (ECU) previously studied using EFOV-US (Adkins et al., 2017; Rakauskas et al., 2023), as well as four additional thumb muscles (APL, EPB, EPL, FPB) not yet measured using this technique. This selection enabled comparisons to the limited existing in vivo literature and allowed for an assessment of how well the imaging methods generalized to both intrinsic (APB, FPB) and extrinsic (APL, EPB, EPL, FPL) thumb muscles. Although ECU is not a thumb muscle, it was included to enable external evaluation of our EFOV-US protocol, as Adkins et al. (2017) only reports fascicle lengths for ECU and Rakauskas et al. (2023) is prior work from our team.

2.4. Image Acquisition

EFOV-US images were captured with a SuperSonic Imagine Mach 30 (Aix-en-Provence, France) by slowly moving the probe along the muscle path from proximal to distal at a constant speed to create a panoramic view of the entire muscle belly. Due to muscle size differences, extrinsic muscles were imaged with the L18–5 probe, while intrinsic muscles were imaged with the LH20–6 probe. Note, the LH20–6 probe is smaller, with lower depth penetration and higher resolution than the L18–5 probe. All images were captured by a single, trained sonographer to ensure consistency.

During imaging, participants were seated in front of the ultrasound machine with their tested arm supported on a table. Arm posture varied depending on the muscle being imaged (Fig. 1). For all muscles except ECU, the participant’s forearm was oriented directly toward the sonographer, with the forearm supinated for FPL, APB, and FPB and pronated for APL, EPL, and EPB. For ECU, the participant’s forearm was pronated, with the elbow flexed and forearm rotated to expose the ulnar side of the arm. Participants were instructed to maintain a relaxed arm posture by sitting in a position that did not place strain on the shoulder, elbow, or wrist joints. Most participants adopted a neutral shoulder and wrist posture with an elbow flexion angle between 120 and 140 degrees. They were also instructed to avoid any voluntary muscle contractions during imaging. Muscle paths were confirmed prior to imaging by observing muscle contractions during standard, muscle-specific movements. Each muscle was imaged multiple times (Fig. 2). Specifically, eight images were taken of each extrinsic muscle (FPL, APL, EPL, EPB, ECU) and 15 images of each intrinsic muscle (APB, FPB).

Figure 1:

Imaging set-up for (A) flexor pollicis longus (FPL), (B) abductor pollicis longus (APL), extensor pollicis longus (EPL), extensor pollicis brevis (EPB), (C) abductor pollicis brevis (APB), flexor pollicis brevis (FPB), and (D) extensor carpi ulnaris (ECU). The arrows indicate the direction the probe was moved during data collection. The thumb was in a relaxed posture for imaging of all muscles.

Figure 2:

Down sampling procedure of ultrasound images from extrinsic (left) and intrinsic (right) muscles. Ultrasound images were taken of each muscle. The best images were determined based on image quality. Ten fascicle lengths were obtained for each muscle.

2.5. Image Analysis & Fascicle Length Measurement

Fascicle length measurements were obtained from a subset of the highest quality images (Fig. 2). To identify high-quality images for analysis, all recorded images were qualitatively assessed based on overall clarity, presence of bright muscle fascicles and fascia, and lack of image distortions. A good-quality image was one where the sonographer was able identify the muscle as a hypoechoic (dark) shape with hyperechoic (bright) muscle borders (Adkins & Murray, 2020), with visible fascicles that smoothly connected the superficial and deep muscle borders (Fig. 3).

Figure 3:

Representative ultrasound images of each muscle with clear muscle borders and visible fascicles. The green-shaded regions highlight the muscle, while the yellow lines indicate traced fascicles. Images are organized by muscle location [top: extrinsic (FPL, APL, EPL, EPB); middle: intrinsic (APB, FPB); bottom: reference (ECU)]. APL and EPL muscles can be imaged together, so the same ultrasound image is used to represent both muscles.

Selected images were imported into ImageJ as DICOM files using the Bio-Formats extension package (Schneider et al., 2012). Image brightness was enhanced using the sharpen tool with default settings to improve visibility of fascicles and muscle borders. Fascicle length was measured as the distance between muscle borders with the segmented line tool. While EFOV-US allows visualization of the entire muscle belly within a single panoramic image, individual fascicles are only continuously visible throughout the entire image if they are aligned with the probe’s imaging plane. Even slight deviations in probe angle or orientation can cause portions of the fascicle to move out of plane and disappear from the image. Thus, only one or two continuously visible fascicles were measured per image, with measured fascicles being selected based on clarity. For extrinsic muscles, which are larger and more likely to contain multiple visible fascicles in a single image, two fascicle lengths were measured from five high quality images for each participant. For intrinsic muscles, one fascicle length was measured from ten high quality images for each participant. This approach resulted in the same total measurements for each muscle [180 measurements per muscle: 2 measurements × 5 images × 18 participants (extrinsic muscles); 1 measurement × 10 images × 18 participants (intrinsic muscles)].

2.6. Statistical Analysis

To summarize the normative fascicle length data measured in this study, mean and standard deviations were calculated for each muscle within each participant and across all participants. Coefficients of variation were also calculated for each muscle in each participant and then averaged across participants to quantify the repeatability of the measured data.

To test the hypothesis that measured fascicle lengths would be more closely aligned with prior EFOV-US data than with cadaveric measurements, two-sided t-tests were performed in GraphPad. These t-tests independently tested whether the measured fascicle lengths were significantly different from those previously reported. For cadaveric literature, statistical comparison was limited to prior work that reported variance metrics (standard deviation or standard error) (Jacobson et al., 1992; Lieber et al., 1990), while additional studies were included for qualitative comparison (Amis et al., 1979; Brand et al., 1981; Cutts et al., 1991; Ruggiero et al., 2016). For in vivo literature, statistical comparison was limited to two thumb muscles (FPL, APB) previously measured by our team (Rakauskas et al., 2023). ECU was included to supplement these comparisons with upper limb in vivo fascicle length data measured by an external group (Adkins et al., 2017). To evaluate relationships between fascicle lengths and anthropometric measurements within participants, regressions were performed and statistically compared using mixed-effects models. Anthropometric measurements were included as fixed effects with participant included as a random effect. An anthropometric measurement was considered a potential predictor of fascicle length if it was statistically significant. The significance level was set as α = 0.05 for all analyses.

3. RESULTS

Precise fascicle length measurements were obtained for each of the muscles (Table 1), as indicated by the small standard deviations of ±0.4 to ±0.8 cm across 18 participants. Coefficients of variation demonstrated high repeatability for the measured thumb muscles: 0.05 for FPL, 0.08 for APL, 0.06 for EPL, 0.08 for EPB, 0.04 for APB, and 0.06 for FPB. Across all participants, the mean thumb muscle fascicle lengths (± SD) were 6.5 ± 0.8 cm for FPL, 3.8 ± 0.4 cm for APL, 4.7 ± 0.5 cm for EPL, 3.7 ± 0.5 cm for EPB, 4.5 ± 0.5 cm for APB, and 3.6 ± 0.4 cm for FPB (Fig. 4A). As expected, measurements from four thumb muscles were statistically different from some values reported in cadaver literature (FPL: Lieber et al., 1992, p < 0.0001; APL: Jacobson et al., 1992, p < 0.0001; EPB: Jacobson et al., 1992, p < 0.0001; FPB: Jacobson et al., 1992, p = 0.0029) (Fig. 4B). Of the two thumb muscles previously studied in vivo, FPL fascicle lengths were similar to those reported by Rakauskas et al. (2023), while APB fascicle lengths were significantly shorter (p = 0.0045) (Fig. 5). ECU fascicle lengths were similar to those reported by Rakauskas et al. (2023) and the novice sonographer in Adkins et al. (2017). However, they were significantly longer than those reported by an expert sonographer who imaged the ECU in a contracted posture (Adkins et al., 2017) (p < 0.0001) (Fig. 5).

Table 1.

Individual Subject Fascicle Length Measurements (in cm)

	Flexor pollicis longus (FPL)
	M1	M2	M3	M4	M5	M6	M7	M8	M9	M10	F1	F2	F3	F4	F5	F6	F7	F8
1	5.63	6.08	6.69	6.32	7.06	7.16	6.25	6.58	7.51	8.15	6.54	6.27	5.88	6.71	6.44	7.15	5.13	5.54
2	5.45	5.82	5.69	7.18	6.40	7.89	5.65	7.27	7.46	8.51	5.97	6.45	6.39	7.22	6.83	7.05	5.60	5.43
3	6.16	5.89	6.75	5.81	7.65	7.20	7.00	6.52	6.59	8.03	5.92	6.26	6.04	6.75	7.18	7.29	5.09	5.18
4	6.46	5.80	5.99	6.47	7.32	7.11	6.48	7.35	6.98	8.19	5.83	6.13	6.00	7.01	7.18	6.62	5.10	5.87
5	6.41	5.71	6.89	5.86	7.66	7.03	7.16	6.83	6.62	7.51	5.73	6.28	5.82	7.23	6.54	6.60	5.01	5.86
6	6.51	5.32	6.66	6.34	7.28	7.66	7.23	7.29	6.65	7.98	6.04	5.98	6.35	6.82	6.69	7.19	5.02	5.32
7	6.00	6.35	6.00	6.52	7.69	7.10	6.86	6.97	7.19	7.14	5.73	6.24	5.80	6.87	6.84	6.59	5.04	5.89
8	6.00	6.15	5.88	6.80	7.96	7.09	6.86	6.74	7.25	7.33	5.92	6.02	6.57	7.32	6.76	6.61	4.89	5.89
9	5.54	6.25	6.26	6.82	6.91	6.62	7.15	6.10	7.39	7.65	5.78	6.09	5.81	6.97	7.02	6.66	4.83	5.54
10	6.26	5.98	5.61	7.14	6.60	7.27	7.34	6.03	7.13	7.66	5.77	6.22	6.23	6.72	7.16	7.36	4.87	5.31
Avg	6.01	5.93	6.24	6.53	7.25	7.21	6.80	6.77	7.08	7.81	5.92	6.19	6.09	6.96	6.86	6.91	5.06	5.58
SD	0.37	0.28	0.45	0.45	0.48	0.33	0.50	0.45	0.33	0.40	0.23	0.13	0.26	0.21	0.25	0.31	0.21	0.26
	Abductor pollicis longus (APL)
	M1	M2	M3	M4	M5	M6	M7	M8	M9	M10	F1	F2	F3	F4	F5	F6	F7	F8
1	5.71	4.26	3.57	4.19	3.67	3.77	3.89	3.80	4.43	3.66	3.44	3.67	3.67	4.45	3.83	4.38	3.71	3.63
2	4.16	4.08	4.20	4.33	3.75	3.51	3.57	4.30	4.10	4.05	3.94	3.35	3.99	3.73	3.93	4.27	3.58	3.44
3	4.26	4.44	3.85	3.91	3.54	3.30	3.25	4.08	4.45	4.65	3.88	3.58	3.47	3.93	3.89	4.42	4.03	3.35
4	4.10	4.53	3.80	3.83	3.43	3.41	3.28	3.74	3.84	3.97	3.25	3.66	3.71	4.09	4.17	4.38	3.67	3.50
5	3.74	4.07	3.58	3.60	3.63	3.90	3.16	4.06	4.19	3.62	3.25	3.75	3.56	3.68	3.99	3.15	3.73	3.41
6	3.66	4.03	3.41	3.62	3.73	4.02	3.02	3.69	3.52	3.77	3.34	3.83	3.69	3.71	3.87	3.88	3.27	3.93
7	5.55	4.43	3.73	4.29	4.04	4.36	3.48	3.92	4.06	3.75	3.51	3.79	2.97	4.48	3.98	3.51	3.85	3.52
8	5.30	4.38	4.23	3.54	3.85	4.28	3.30	3.74	3.15	3.93	3.28	3.19	3.18	4.76	3.92	3.80	3.53	3.06
9	4.63	4.27	4.13	3.50	3.95	3.54	4.10	3.35	3.99	3.81	3.33	3.03	3.25	3.69	3.75	4.40	4.26	3.57
10	4.50	4.60	3.95	3.89	3.88	3.39	3.70	3.48	3.88	3.68	3.05	3.50	3.07	3.62	4.39	3.80	3.41	3.79
Avg	4.56	4.30	3.84	3.87	3.70	3.75	3.47	3.81	3.96	3.89	3.43	3.54	3.50	4.01	3.97	4.00	3.70	3.52
SD	0.69	0.20	0.27	0.29	0.20	0.36	0.32	0.27	0.38	0.29	0.27	0.25	0.30	0.39	0.18	0.42	0.30	0.23
	Extensor pollicis longus (EPL)
	M1	M2	M3	M4	M5	M6	M7	M8	M9	M10	F1	F2	F3	F4	F5	F6	F7	F8
1	5.39	4.17	4.87	5.21	4.85	5.20	5.15	4.89	5.38	4.94	4.36	4.19	5.37	5.40	4.15	5.40	3.66	3.32
2	5.16	4.05	4.68	5.36	4.54	5.21	5.28	4.62	5.17	4.83	4.15	4.63	5.30	5.01	4.11	4.33	4.02	3.50
3	5.00	4.04	4.92	4.75	5.01	4.92	5.07	5.25	4.50	5.06	4.35	4.57	5.21	5.12	4.58	4.97	4.22	3.64
4	4.57	4.49	4.83	4.86	5.30	5.02	4.89	5.13	5.01	4.74	3.94	5.17	5.11	5.14	4.67	4.70	4.49	3.52
5	5.06	4.02	4.56	5.17	4.89	4.56	4.93	4.97	5.53	5.25	4.22	4.45	5.35	5.27	4.48	3.82	4.30	3.28
6	4.49	4.12	5.10	4.87	5.19	4.69	5.13	4.55	5.53	5.19	4.39	4.54	4.90	5.50	4.36	3.46	4.01	3.64
7	4.43	4.37	4.53	4.80	4.71	5.14	5.29	4.87	5.62	5.08	4.45	5.47	5.45	5.07	4.71	3.85	4.18	3.36
8	5.14	4.49	4.67	5.19	4.27	5.05	5.26	5.15	5.51	4.90	4.51	4.96	5.68	4.73	4.56	3.98	4.03	3.47
9	5.11	4.14	4.64	4.93	5.21	5.18	4.95	4.92	4.82	5.07	4.34	4.74	4.80	5.20	4.65	4.65	4.09	3.14
10	4.25	4.01	4.86	5.04	5.29	4.65	5.22	5.25	5.42	4.91	4.52	4.36	4.74	5.66	4.68	4.21	3.44	3.81
Avg	4.86	4.19	4.77	5.02	4.93	4.96	5.12	4.96	5.25	5.00	4.32	4.71	5.19	5.21	4.49	4.34	4.04	3.47
SD	0.37	0.18	0.17	0.20	0.33	0.23	0.14	0.23	0.35	0.15	0.17	0.37	0.29	0.25	0.21	0.56	0.29	0.19
	Extensor pollicis brevis (EPB)
	M1	M2	M3	M4	M5	M6	M7	M8	M9	M10	F1	F2	F3	F4	F5	F6	F7	F8
1	4.15	4.42	3.07	3.45	3.97	4.68	3.58	3.82	3.35	3.83	4.06	3.72	4.78	4.48	3.23	3.75	2.99	2.19
2	4.16	4.27	3.57	3.75	3.77	4.30	3.59	4.45	3.70	3.46	3.36	3.63	4.75	4.18	2.91	3.17	2.65	2.66
3	3.98	4.79	3.51	3.72	4.02	3.43	3.63	3.92	3.38	3.84	3.92	3.98	4.87	3.66	4.19	3.84	3.36	2.89
4	3.91	4.70	3.12	3.65	3.91	3.63	3.61	3.80	3.33	4.00	3.49	3.42	4.70	3.82	3.88	4.11	3.03	2.84
5	3.31	4.69	3.73	3.90	4.58	3.98	3.47	4.05	3.28	3.92	4.75	3.39	3.97	3.90	3.22	3.04	3.25	2.79
6	3.43	4.46	3.40	3.82	4.34	4.14	4.17	3.74	3.26	3.74	4.92	3.21	4.81	3.62	3.53	3.07	2.72	2.76
7	3.12	4.50	3.05	3.70	4.59	3.77	4.15	3.69	3.65	4.01	3.91	3.49	4.54	4.22	3.63	3.01	3.22	2.95
8	3.14	4.41	2.94	3.45	3.96	3.53	3.51	3.82	3.53	3.84	3.92	3.93	4.48	3.67	3.47	3.32	2.97	3.27
9	3.16	4.82	3.15	4.07	4.08	4.21	3.98	4.09	3.35	3.96	3.62	4.00	4.70	3.49	3.68	3.06	2.61	3.51
10	3.13	4.63	3.30	3.37	3.79	3.88	3.85	3.76	3.23	3.91	3.64	3.84	4.78	3.60	3.86	3.10	3.30	3.11
Avg	3.55	4.57	3.28	3.69	4.10	3.96	3.75	3.91	3.40	3.85	3.96	3.66	4.64	3.86	3.56	3.35	3.01	2.90
SD	0.42	0.17	0.25	0.21	0.29	0.37	0.25	0.22	0.16	0.15	0.49	0.26	0.25	0.31	0.36	0.38	0.26	0.34
	Abductor pollicis brevis (APB)
	M1	M2	M3	M4	M5	M6	M7	M8	M9	M10	F1	F2	F3	F4	F5	F6	F7	F8
1	5.54	4.25	4.72	4.69	4.67	5.12	4.97	4.15	5.19	4.89	3.75	4.19	4.14	3.85	4.10	3.70	3.72	4.34
2	5.46	4.43	4.73	4.54	4.46	4.79	5.17	4.66	4.97	4.68	4.04	4.03	4.11	4.28	3.94	4.23	4.11	4.10
3	5.51	3.89	4.56	4.52	4.48	4.97	5.34	4.53	5.25	4.74	4.13	3.73	3.99	4.10	4.06	4.17	3.73	4.08
4	5.43	4.39	4.40	4.92	4.77	4.61	5.13	4.48	4.96	4.72	3.70	4.30	4.08	4.06	3.82	4.16	3.60	3.85
5	5.28	4.32	4.50	4.66	4.53	4.68	5.36	5.07	4.81	5.01	3.83	4.06	4.23	4.36	4.43	3.89	3.83	4.20
6	5.56	4.29	4.30	4.63	4.69	4.94	4.51	5.13	4.91	4.80	4.15	4.11	4.01	4.35	4.17	3.71	3.79	3.55
7	5.54	4.06	4.67	4.73	4.44	4.62	4.83	4.60	5.28	4.75	3.80	3.89	4.37	4.16	3.80	3.89	4.01	4.15
8	5.48	3.85	4.83	4.98	4.61	4.71	4.56	4.34	4.97	5.03	3.73	3.97	4.08	4.22	4.45	3.94	4.08	3.77
9	5.50	4.00	4.98	4.84	4.60	4.64	5.15	4.63	4.62	5.13	3.75	4.07	4.13	3.58	4.29	4.25	4.10	3.79
10	5.50	3.92	5.02	4.95	4.62	4.71	4.70	4.72	5.17	4.96	4.12	4.15	3.99	3.98	3.91	4.14	4.10	4.35
Avg	5.48	4.14	4.67	4.75	4.59	4.78	4.97	4.63	5.01	4.87	3.90	4.05	4.11	4.09	4.10	4.01	3.91	4.02
SD	0.08	0.21	0.22	0.16	0.10	0.16	0.29	0.28	0.20	0.15	0.18	0.15	0.11	0.23	0.22	0.20	0.18	0.25
	Flexor pollicis brevis (FPB)
	M1	M2	M3	M4	M5	M6	M7	M8	M9	M10	F1	F2	F3	F4	F5	F6	F7	F8
1	4.17	3.82	3.60	3.38	3.13	3.51	3.66	3.40	3.57	3.59	3.13	3.49	3.23	3.55	3.74	3.90	3.13	3.53
2	4.30	3.89	3.32	3.46	3.20	3.38	3.89	3.61	3.91	3.34	3.33	3.16	3.15	3.62	3.62	3.84	2.97	3.37
3	4.05	3.47	3.35	3.93	3.79	3.87	4.01	3.50	3.50	3.50	3.56	3.21	2.95	3.39	3.57	4.17	2.83	3.19
4	4.42	4.01	3.72	3.29	3.54	3.73	4.37	3.49	3.63	3.70	3.22	3.11	3.06	3.32	3.63	3.87	3.00	3.21
5	4.27	4.02	3.50	4.25	3.98	3.93	4.25	3.35	4.46	3.08	3.59	3.48	3.14	3.64	3.43	3.84	3.00	3.42
6	4.20	3.89	3.91	3.79	3.45	3.65	4.20	3.45	4.27	3.07	3.26	3.15	3.34	3.17	3.37	3.78	3.25	2.59
7	4.29	3.64	3.58	3.15	3.95	3.88	3.69	3.36	3.77	3.38	3.68	3.12	3.14	3.16	3.36	3.80	3.18	3.58
8	4.28	3.75	3.96	4.26	3.58	4.45	4.16	3.43	4.06	3.67	3.23	3.01	3.33	3.25	3.42	3.41	2.92	3.84
9	4.04	3.81	4.17	4.07	3.23	4.10	3.52	3.43	3.54	3.78	3.21	2.78	3.22	3.15	3.13	3.66	2.95	3.55
10	4.02	3.73	3.38	3.78	3.44	4.30	3.61	3.80	3.57	3.51	3.27	3.27	3.31	3.30	3.20	4.06	3.17	3.26
Avg	4.21	3.80	3.65	3.73	3.53	3.88	3.94	3.48	3.83	3.46	3.35	3.18	3.19	3.36	3.45	3.83	3.04	3.35
SD	0.13	0.16	0.27	0.38	0.29	0.32	0.29	0.13	0.32	0.24	0.18	0.20	0.12	0.18	0.18	0.20	0.13	0.32
	Extensor carpi ulnaris (ECU)
	M1	M2	M3	M4	M5	M6	M7	M8	M9	M10	F1	F2	F3	F4	F5	F6	F7	F8
1	4.44	4.73	4.19	4.58	4.62	4.42	4.57	4.73	4.29	4.24	4.32	3.64	3.85	3.66	4.16	4.36	3.97	3.01
2	4.57	4.48	4.37	4.26	4.77	4.81	4.40	4.71	4.28	4.50	3.55	3.52	3.35	3.96	3.74	4.26	3.68	3.04
3	4.92	4.37	3.98	4.94	4.55	4.46	4.39	4.35	4.53	4.36	3.76	3.63	3.55	3.56	3.66	4.23	3.92	3.55
4	4.94	4.39	4.28	4.58	4.46	4.73	4.84	4.36	4.37	4.48	3.45	3.98	4.16	3.48	3.74	4.08	3.76	3.25
5	5.21	4.46	3.96	5.08	4.24	4.75	4.43	4.25	4.78	4.22	4.17	3.58	3.07	3.73	3.65	4.01	4.17	3.40
6	4.74	4.12	4.17	5.29	4.74	4.82	4.34	4.35	4.79	4.06	3.81	3.49	3.44	3.56	3.64	3.74	3.97	3.04
7	5.11	4.23	3.46	4.64	4.99	5.02	4.47	4.56	4.11	4.15	3.70	3.86	3.46	3.90	4.15	4.04	3.63	3.48
8	4.52	4.62	3.52	4.78	4.78	4.78	4.46	4.40	4.51	4.28	3.69	3.98	3.33	3.44	4.12	3.88	4.15	3.25
9	5.06	4.75	3.76	4.65	4.53	5.06	4.42	4.47	4.55	4.36	4.00	3.95	3.85	3.97	4.42	4.27	3.97	3.80
10	4.61	4.18	4.03	4.80	4.59	5.15	4.91	4.62	4.78	4.61	3.79	3.60	3.89	3.46	4.00	3.80	3.89	3.22
Avg	4.81	4.43	3.97	4.76	4.63	4.80	4.52	4.48	4.50	4.33	3.82	3.72	3.59	3.67	3.93	4.07	3.91	3.30
SD	0.26	0.21	0.29	0.28	0.19	0.22	0.19	0.16	0.23	0.16	0.26	0.19	0.31	0.20	0.26	0.20	0.17	0.24

Abbreviations: M = Male, F = Female, Avg = Average, SD = Standard Deviation

Figure 4:

A) Total average fascicle lengths measured (dark shaded bars) compared to B) cadaver values (unfilled bars) reported in the literature. Error bars indicate standard deviation. Asterisks indicate statistically significant differences (*p<0.01, **p<0.0001). Note, Amis et al. (1979), Brand et al. (1981), Cutts et al. (1991), and Ruggiero et al. (2016) did not report standard deviations across participants in their results; thus, statistical comparisons with these studies could not be performed.

Figure 5:

Total average fascicle lengths measured (dark shaded bars) compared to EFOV-US values (light shaded bars) reported in the literature. Error bars represent standard deviation. Asterisks indicate statistically significant differences (*p<0.01, **p<0.0001).

Significant relationships were observed between fascicle lengths and various anthropometric measurements (Fig. 6). Among the extrinsic thumb muscles (FPL, APL, EPL, EPB), height, forearm length, hand length, and thumb length were significantly associated with only FPL fascicle lengths. Weight was significantly correlated with APL fascicle lengths. No significant relationships were found for EPL or EPB fascicle lengths. For the intrinsic muscles (APB, FPB), fascicle lengths were significantly associated with height, forearm length, hand length, hand width, and thumb length. Additionally, weight was significantly correlated with FPB fascicle lengths.

Figure 6:

Relationship between average fascicle lengths for the flexor pollicis longus (FPL), abductor pollicis longus (APL), extensor pollicis longus (EPL), extensor pollicis brevis (EPB), abductor pollicis brevis (APB), and flexor pollicis brevis (FPB) with anthropometric measurements. Each point represents a participant average (circles: males; triangles: females), regression lines examine the relationship across participants, and the shaded regions represent a 95% confidence interval of the regression. Significant regressions are depicted by the inclusion of p- and R²-values.

4. DISCUSSION

This study establishes normative data on thumb muscle fascicle lengths in a young, healthy population using EFOV-US imaging. Measurement consistency is indicated by the small standard deviations within (±0.1 to ±0.7 cm) and across (±0.4 to ±0.8 cm) participants. The coefficients of variation observed in this study (range: 0.04–0.08) were generally comparable to those calculated from prior cadaveric literature (range: 0.05–0.14) (Jacobson et al., 1992; Lieber et al., 1990), with equal values observed for FPL and EPL and slightly lower values observed for APL, EPB, APB, and FPB in this dataset. Compared to in vivo literature (range: 0.04–0.07) (Rakauskas et al., 2023), the coefficients of variation from this study were equal (APB) and slightly lower (FPL). In this context, the data herein fill a critical gap in musculoskeletal research, where the thumb has remained understudied. By providing normative fascicle length data for six thumb muscles, this work supports more accurate modeling of thumb biomechanics and lays the foundation for identifying muscular deviations linked to thumb pathology.

Differences in fascicle lengths between this study and previous in vivo literature likely stem from variations in imaging posture (Adkins et al., 2017; Rakauskas et al., 2023). In Adkins et al. (2017), two sonographers collected EFOV-US images, and significant differences were observed between the measurements obtained by the expert sonographer and those reported in this study. The expert sonographer positioned participants with their elbow at 90° flexion and wrist ulnarly deviated at 30°, which shortens the ECU. The novice sonographer positioned participants with their elbow at 90° flexion, wrist at neutral, and fingers relaxed. No statistical differences were found between the novice sonographer’s measurements and those in this study. Rakauskas et al. (2023) also standardized participant posture for APB imaging by positioning participants with the shoulder at 90° flexion, elbow fully extended, and forearm supinated at the start of imaging. Differences in forearm supination in this study may have affected measurements, as some male participants had difficulty achieving full supination for imaging of the FPL, APB, and FPB muscles. To maximize comfort during the 2–3-hour imaging session, participants were instructed to remain relaxed. Prioritizing comfort helped reduce unintended movements (e.g., fidgeting, shifting) that could arise from forced positioning, thereby minimizing movement artifacts and ensuring image quality. However, variability in relaxation levels may have introduced differences in muscle conditions across participants. Despite posture-related differences, the fascicle length data collected in this study remain highly valuable, as they were acquired in relaxed, natural postures that reflect real-world muscle conditions and provide the most extensive in vivo data for thumb muscles to date.

Differences in fascicle lengths between the measurements in this study and those reported in cadaveric literature are likely attributed to differences in living and preserved tissues (Cutts, 1988; Holzbaur et al., 2007). All cadaveric studies referenced here obtained fascicle length measurements through muscle dissection (Amis et al., 1979; Brand et al., 1981; Cutts et al., 1991; Jacobson et al., 1992; Lieber et al., 1992; Ruggiero et al., 2016). Among these, two studies (Amis et al., 1979; Lieber et al., 1992) reported embalming cadavers with fingers in a flexed position. This fixation technique may have passively shortened the muscles and resulted in shorter fascicle lengths compared to those measured in this study. In contrast, other studies (Brand et al., 1981; Cutts, 1988; Jacobson et al., 1992) tended to report longer fascicle lengths than those measured in this study. During dissection, muscles are often manipulated to expose and separate them from surrounding tissues. Additionally, some studies remove connective tissues before taking fascicle measurements. These connective tissues provide structural support to the muscle and contribute to the overall passive resistance to lengthening. Loss of these tissues along with the dissection process could inadvertently stretch the muscle fascicles.

Significant regressions were found between the fascicle lengths of one extrinsic muscle (FPL) and two intrinsic muscles (APB, FPB) and various anthropometric measurements. These results differ from those of Rakauskas et al. (2023), who reported a significant relationship only between APB fascicle length and hand length. A likely explanation for this discrepancy is sample size. Rakauskas et al. (2023) included just eight participants, whereas the current study examined 18 participants. A larger sample improves statistical power and increases the likelihood of detecting true associations. It is possible that the limited variability in body size in the smaller sample reduced sensitivity to other relationships. Together, these studies highlight the importance of continued data collection to validate and refine anthropometric models for estimating muscle architecture. By expanding the sample size, it may become feasible to develop regression models based on simple, easily obtainable anthropometric measurements. Such models could complement normative datasets by offering a more efficient way to estimate fascicle lengths in broader populations.

All fascicle measurements were manually traced, which is a time-consuming process that takes approximately 1 hour per muscle. While automated methods for measuring fascicle length have been established for lower limb muscles (Farris & Lichtwark, 2016; Kilpatrick et al., 2023), their applicability to upper limb muscles is limited. Farris & Lichtwark (2016) used B-mode ultrasound videos to track fascicle length during movement. This method requires the entire fascicle to remain within the probe’s field-of-view throughout the movement, which is not feasible in the thumb where muscles often exceed the probe’s visible range. Separately, Kilpatrick et al. (2023) developed an ultrasound analysis method to estimate fascicle length, pennation angle, and curvature using static B-mode images of the tibialis anterior. The tibialis anterior is a superficial muscle that is not surrounded by many overlapping structures, which allows for clear visualization of fascicles and high image contrast. This is not the case in the thumb, where muscles are smaller, deeper, and closely packed together. These anatomical differences make fascicle identification and segmentation in B-mode images more challenging. As a result, manual measurements remain necessary for upper limb fascicle length analysis. However, the manually traced images from this study could serve as a training dataset for developing upper limb–specific machine learning models. Such models could reduce analysis time, improve measurement consistency, and facilitate broader use of ultrasound-based muscle architecture analysis in both research and clinical settings.

This study is limited by the small sample size. Future work should expand this analysis to larger cohorts, including healthy, older adults and individuals with musculoskeletal pathologies in the thumb. This expansion would allow for evaluation of age-related differences in thumb muscle fascicle lengths and healthy versus pathologic comparisons. In this study, the same person acquired and analyzed all ultrasound images. This ensured consistency in image interpretation, as the individual performing the analysis was familiar with how the images were captured, including probe orientation, pressure distribution, and any subject-specific challenges encountered during data collection. This familiarity may reduce variability in measurements compared to datasets analyzed by multiple independent raters. However, future studies should assess inter-rater reliability to determine the generalizability of these findings across different analysts.

Despite these limitations, the normative data presented in this study provide a critical step toward advancing our understanding of thumb muscle architecture. By establishing normative values for fascicle lengths in six thumb muscles, this work supports future efforts in musculoskeletal modeling and identifying pathologies through muscular deviations.