A total of 474 adults were recruited by the Motor Assessment Center of the Collaborative Research Center 1,451 (Fink et al., 2022) of the German Research Foundation (DFG) from April 2021 to October 2023. The following inclusion criteria were applied for the current study: age between 21 and 90 years, non-left-handedness (defined as a laterality quotient larger than −28 as determined by the Edinburgh Handedness Inventory; Oldfield, 1971), the absence of any relevant depressive symptoms (operationalized by unremarkable results in the Beck Depression Inventory [BDI-II, ≤13), (Kühner et al., 2007) and the Montgomery-Åsberg Depression Rating Scale (MADRS ≤12) (Montgomery and Asberg, 1979)], as well as having no prior history of neurological, psychiatric, or orthopedic diseases (selection of participants in Figure 1). The final sample consisted of 444 healthy participants (mean age of 52.5 years ±18.5 years, with a range of 21 to 88 years; 254 [57.2%] females) (descriptive statistics in Supplementary Table S1). Trained examiners administered a standardized motor ability assessment in two laboratories at either the Department of Neurology of the University Hospital of Cologne or the Institute of Neuroscience and Medicine (INM-3 – Cognitive Neuroscience) at the Research Center Jülich. All participants provided written informed consent. The ethics committee of the Medical Faculty of the University of Cologne approved the study performed under the Declaration of Helsinki.

To account for the broad age range and interindividual variability in motor performance, we employed five robust and reliable motor tests to assess individual motor abilities comprehensively (Kwakkel et al., 2019) and a measure of handedness. All tests were administered by three trained study nurses and according to a standardized test protocol including detailed written instructions to be read out to the participants to guarantee comparability. Additionally, there were weekly meetings to discuss occurring questions or uncertainties regarding test administration or interpretation.

Statistical analyses were performed using IBM SPSS Statistics (Statistical Package for the Social Sciences, version 29, SPSS Inc., Chicago, Illinois, United States) and Prism (GraphPad Software, LLC, Version 9.4.1). For all analyses, a significance level of p < 0.05 was set. If not otherwise stated, data are depicted as mean ± standard error of the mean (SEM).

To provide an overview of the motor performance for the entire sample of the 444 non-left-handed healthy participants, we report the results for the basic and complex motor tests as well as the ratios between the non-dominant hand’s performance and dominant hand’s performance collapsed across age groups and sexes in Table 1.

With respect to the tests probing more complex motor functions, the ARAT scores showed no variance. All 444 participants achieved the maximal total score of 57 points. Hence, ARAT was excluded from all further analyses, given the strong ceiling effects of this test in healthy participants, which did not add additional information to the assessment of motor performance.

Pearson correlation analysis was applied for the parameter of a given motor test and age as a continuous variable to assess the motor performance in the different motor tests across the adult life span. There was a significant decline in performance in all basic and complex motor tasks across the examined age range (21–88 years) (Figure 2). For clarity, we only report the results of correlation analyses for the dominant hand in the text. The results of the correlation analyses for the non-dominant hand are documented in Table 2. Grip strength (p < 0.001, r = −0.349) and FTF (p < 0.001, r = −0.369) demonstrated a significant negative correlation with age. In addition, the number of stacked pins in the Purdue Pegboard Test decreased significantly with age (p < 0.001, r = −0.641). As expected, the time required to perform the JTT exhibited a significant positive correlation with age (p < 0.001, r = 0.491).

Notably, the performance ratios of the four basic and complex motor tests (performance with the non-dominant hand divided by the performance of the dominant hand) did not demonstrate any significant correlation with age (all p > 0.999) and thus remained stable across the examined age range (21–88 years, see Table 2).

For the basic motor tests (grip strength, FTF) and the more complex motor tests (Purdue Pegboard Test, JTT), we calculated repeated measures ANOVAs with the within-subject factor HAND (dominant, non-dominant) and the between-subject factors AGE [young (21–40 years), middle-aged (41–60 years), older (61–88 years)] and SEX (men, women). We included the detailed statistics and the post hoc tests in Table 3.

There were significant main effects for the within subject factor HAND and the between subject factor AGE for the basic motor tests (grip strength, FTF) as well as for the complex motor tests (Purdue Pegboard Test, JTT).

Additionally, we observed main effects for the between subject factor SEX for three motor tests (grip strength, FTF, Purdue Pegboard Test). Notably, the JTT revealed only a trend for the factor SEX.

Post hoc comparisons indicated that the performance of the dominant hand was better than that of the non-dominant hand regarding every test.

Additionally, men achieved higher scores than women concerning grip strength and FTF, whereas women stacked more pins per time than men during the Purdue Pegboard Test.

Moreover, young participants outperformed older participants in all scores.

Notably, the pattern of this age-dependent decline differed within the tests. There was a consistent age-dependent reduction regarding Purdue Pegboard Test performance and FTF.

Grip force decreased significantly from young to middle-aged participants. However, there was no further significant decrease for older participants (Figure 3).

JTT, on the other hand, showed stable performance scores for the young and middle-aged groups, but a significant decrease from middle-aged to older participants (see Figure 3 and Table 3 for post hoc tests).

There was a significant interaction of the factors HAND x AGE for the performance in the Purdue Pegboard Test and FTF (see Figure 3 and Table 3 for statistics and post hoc tests).

Regarding the Purdue Pegboard Test, the difference between dominant and non-dominant hands became less pronounced with increasing age.

For FTF, the interaction was driven by a higher difference between dominant and non-dominant hands for the middle-aged.

For the current cohort of 444 healthy participants, we provide descriptive statistics (i.e., mean, median, range, SD, SEM) for the performance in the four motor tests with either the dominant or non-dominant hands in Supplementary Table S2. Based on the repeated measures ANOVA, values are given for six groups of participants comprising young (age between 21 and 40 years) as well as middle-aged (age between 41 and 60 years) and older (age between 61 and 88 years) women and men. These values could be used to compare single-subject data to the respective control group (Crawford et al., 1998).

A multiple linear regression analysis investigated whether motor performance significantly predicted the participants’ age. Grip strength, Purdue Pegboard Test and JTT performances, and FTF were utilized as predictors, while age served as the dependent variable. Indeed, the overall regression model yielded statistically significant results [Adjusted R² = 0.502, F_(8,435) = 56.854, p < 0.001], indicating that the predictors collectively had a significant effect on the prediction of age. Specifically, grip strength (β = −0.290, p = 0.003), performance in the Purdue Pegboard Test (β = −0.462, p < 0.001), and finger-tapping frequency (β = −0.166, p = 0.005) emerged as significant predictors of age.

To disclose potential components underlying the performance in the different basic and complex motor tests, we conducted a PCA on the motor scores of both hands for the entire cohort. This approach identified three principal components, which collectively accounted for 86.2% of the total variance observed. Component 1 (termed motor component “dexterity”), explaining 48.1% of the variance, included strong loadings from the Purdue Pegboard Test and the JTT. The grip strength of both hands loaded on component 2 (termed motor component “force”) accounted for 22.3% of the variance. Finally, component 3 (termed motor component “speed”), explaining 15.8% of the variance, was related to the FTF of both hands. Importantly, none of these tests showed reliable loadings on multiple components. Table 4 provides the specific loadings of each test.

As we have encountered age-dependent differences in performance, we also performed separate PCAs for the age groups (21–40 years, 41–60 years, and 61–88 years). Remarkably, the partition into the three components and their respective assignment remained unchanged (variance explained young: 80.4%, middle-aged: 78.9%, older: 84.8%, see Supplementary Table S3 for details), proving a robust separation into the three fundamental motor components dexterity (JTT and Purdue Pegboard Test), (grip) force, and speed (FTF).

As a next step, we calculated sex-specific PCAs (women and men; see Supplementary Table S4). The patterns observed for the entire cohort were replicated for both groups. The three fundamental motor components dexterity, force, and speed explained 84.7 and 85.4% of the total variance in women and men, respectively.

In summary, the PCA of motor performance parameters exposed a clear separation into three fundamental motor components, which were robust across sex and age.

We extracted individual component scores based on the three motor components (dexterity, force, and speed) identified by the above-described PCA. These scores represent the participant’s relative position along the principal component based on their weighted performance in the respective assessments. The young group, including young women and men, served as the reference. Due to the heterogeneity of motor performance, we aimed to group the participants in an unbiased and data-driven way, using K-means clustering. This approach generated two subgroups of healthy participants without relying on subjective criteria like age or fixed thresholds, appreciating individual strengths and weaknesses of the participants.

The first cluster comprised 148 (33.3%) participants, while the second cluster comprised 296 (66.7%) participants. Within the young group, 13 participants (8.5%) were assigned to the first cluster, whereas 21 (19.8%) of the middle-aged group and 114 (61.6%) of the older group were assigned to the same cluster. One hundred forty young (91.5%), 85 middle-aged (80.2%), and 71 (38.4%) older participants were assigned to the second cluster. Notably, significant differences were observed between these healthy subgroups, indicating a clear separation between high-performing participants (second cluster) and those displaying reduced performance in specific motor tests (first cluster, see Figure 4).

The study replicated the age-related reduction in grip strength, finger tapping speed, and performances in the Purdue Pegboard and the Jebsen-Taylor Hand Function tests. These findings align with prior studies investigating these motor tests separately (Desrosiers et al., 1995b; Aoki and Fukuoka, 2010; Dodds et al., 2014; van der Willik et al., 2020).

Given that aging is a complex process, many factors contribute to a general decline in motor functions. They include the physiological reduction in muscle mass (Soyuer et al., 2023), reduced coordination or neuromuscular impairments (Carson, 2018; Clark, 2019), as well as changes in sensory, visual, and attention capabilities (Heintz Walters et al., 2021).

Importantly, we could show differential, task-specific patterns of motor performance decline with age (see Figure 2).

Both the FTF and the Purdue Pegboard test revealed a consistent performance decline for each age group, which aligns with prior studies (Hubel et al., 2013; Stijic et al., 2023). In contrast, the reduction of grip force was only significant between young (21–40 years) and middle-aged (41–60 years) participants. There was no significant grip force difference between middle-aged and older participants. This finding is in common with Dodds et al. (2014), as well as Gómez-Campos et al. (2022), who described a peak grip force around the age of 40 years and a following decline. Nevertheless, another pattern appeared for the JTT, for which the performance was stable in the groups of young and middle-aged participants. The JTT performance reduction became significant when comparing the middle-aged and older groups. A potential reason for this pattern might be the resemblance of the JTT subtasks with activities of daily living (ADL), which are limited in older participants only.

Remarkably, our investigation revealed that the performance ratios between the non-dominant and dominant hands did not exhibit any relevant age-dependent changes. Across all age groups, the dominant hand consistently outperformed the non-dominant hand. This finding aligns with the findings of Crosby et al. (1994), who obtained comparable grip strength ratios in 214 participants with an age range of 16–63 years. Additionally, we could show that the hand performance ratios were similar across the different tests, indicating a high reliability of the hand performance ratios as a potential robust metric of proper hand function.

Although performance asymmetries, to a certain extent, do not need to indicate abnormalities (Pool et al., 2014), increased asymmetry indices of hand function can potentially serve as an early indicator of evolving motor deficits. For instance, Parker et al. (2021) associated hand-grip strength asymmetry of more than 20% with an increased likelihood of future limitations in instrumental activities of daily living (IADL) in 18,235 Americans older than 50 years. This finding is further supported by research linking asymmetries of motor functions with functional limitations (Collins et al., 2020), morbidity (Klawitter et al., 2022), risk of future falls (McGrath et al., 2021), and even cognitive deficits (Jia et al., 2023).

Moreover, we observed a significant interaction of HAND (dominant versus non-dominant hand) and AGE for the Purdue Pegboard Test performance. At generally lower performance levels, older participants (61–88 years) showed less pronounced differences between their dominant and non-dominant hand performances compared to middle-aged (41–60 years) and young participants (21–40 years) (Figure 3). These findings confirm but also extend (in a different age range) previous results in a sample of 98 right-handed participants ranging from 5 to 24 years (Roy et al., 2003). These authors also showed “a right-hand advantage in performance which was larger in the younger than the older participants” in the Annett pegboard task.

Notably, there was an additional interaction of HAND and AGE regarding FTF. While young and older participants showed similar differences between the finger-tapping frequencies of the dominant and non-dominant hands, this difference was more pronounced in the middle-aged group. This pattern could be potentially explained by a differential age-related decrease in the finger-tapping performance of the two hands. If the decline of finger-tapping frequency of the non-dominant hand starts earlier (already in the middle-aged group) than that of the dominant hand (present in the older group), then the difference in finger tapping between the non-dominant and dominant hands would be maximal in the middle-aged group.

Thus, the previous and our current findings underscore the importance of examining absolute hand performance scores and considering the relative performance between the dominant and non-dominant hands as valuable research and potential clinical parameters.

Regarding the robust correlation between motor performance and age, we conducted a linear regression analysis, utilizing the motor scores to predict age. Grip strength, Purdue Pegboard Test performance, and FTF significantly predicted age, underlining the close interplay of different motor abilities. Consistent with our results, Martin et al. (2015) described an increasing association between strength and dexterity, which gets more pronounced with age. Pérez-Parra et al. (2023) performed multiple linear regression analyses associating grip strength with the JTT, the Nine-Hole Peg test, and aging. Notably, their findings align with our PCA results, as every motor component (dexterity, force, and speed) was significantly represented in our regression analysis, emphasizing the importance of a multidimensional approach to assess motor function.

With an R²-value of 0.502, our linear regression model still contains relevant unexplained variance. One reason might be the age-dependent increasing variability caused by “high-and low-performing” participants’ as described in the section on K-means clustering. That performance variability increases with age has been shown for finger-tapping frequency (Hubel et al., 2013). Such an age-dependent variability-increase challenges a linear regression model, leading to more unexplained variance, i.e., reducing the predictive power of the linear regressions.

The current comprehensive multidimensional assessment of motor function uncovered sex-dependent variations in performance across the different motor tests. Specifically, men exhibited significantly higher grip strengths and finger-tapping-frequencies, while women performed significantly better in the Purdue Pegboard Test. This pattern of results aligns with the 1,000 Norms project, which suggested that men performed better in gross motor tasks while females outperformed men in dexterous tasks (McKay et al., 2017a,b). Similarly, studies claimed that men are faster in repeating a single movement, while women outperform men in the speed of programming a sequence of hand movements (Nicholson and Kimura, 1996; Schmidt et al., 2000). It has been suggested that these performance patterns may be associated with testosterone-mediated fast-twitch muscle fibers or morphological asymmetry patterns in the brain, which were mainly described for the cerebellum (Fan et al., 2010; Hubel et al., 2013; Shigehara et al., 2022), an issue that warrants further investigation.

Interestingly, there was no significant sex effect regarding the JTT. Different strategies can be used to accomplish the JTT subtasks, as suggested by kinematic analysis during JTT performance (Kontson et al., 2020). Thus, men and women might use different strategies that, nevertheless, result in similar JTT performance.

Principal component analysis (PCA) revealed three motor components – namely dexterity, force, and speed - persisting throughout the entire lifespan and regardless of sex. Notably, the basic motor assessments grip strength and FTF separated into two independent motor components, while the complex tasks Purdue Pegboard test and JTT formed one joint motor component. The separation of the different motor tasks was data-driven and did not depend on artificial classifications of motor assessments.

The three robust motor components underline the importance of considering different motor functions when assessing motor performance, as they prove the existence of different and independent motor abilities. In this vein, Bowden and McNulty already showed, that strength, speed, and dexterity need to be evaluated independently (Bowden and McNulty, 2013). Yet, they did not show whether the motor tasks represented different and independent motor functions.

In line with this, Ingram et al. (2019) introduced a physiological profile assessment (PPA), a multimodal test battery to assess independent motor functions. It measured grip strength, reaction time, dexterity, tactile sensation, bimanual coordination, and a functional task. Putting these tests in relation to each other, the authors described four domains of upper limb function: gross motor skills, arm stability, fine motor control, and tactile discrimination (Ingram et al., 2023). Our study confirms and extends these previous findings by demonstrating that the separation of motor components remains independent of age and sex.

This allows for an analysis of motor function across the whole population through the comparison of motor components.

In contrast to Ingram et al. (2019), who calculated a composite score from the average motor performances to discriminate between healthy and impaired participants, we used a data-driven K-means-clustering approach. This approach was independent of previously built hypothesis or cut-off values. K-means clustering based on the three motor components revealed those participants in the cohort whose motor performance was disproportionally low. Within each age group, it successfully segregated individuals into two distinct subgroups with overall low (first group) or high (second group) performance levels in the applied tests of motor functions.

Thus, the importance of a multidimensional assessment of motor function is also highlighted by the results of the K-means clustering approach. For example, participants who performed well in force or speed could still be assigned to the “low-performing” group due to substantially reduced dexterity. The current approach enables the discrimination between orthogonal agers, who maintain their motor abilities in advanced age, and disproportional agers, who show a pronounced reduction in motor function. Importantly, this approach does not define pathological behavior. However, recognizing a disproportionate reduction in upper limb motor function (i.e., over and above the reduction that is expected by that person’s age) could serve as an early indicator of abnormal aging. Therefore, K-means clustering by motor performance might help to detect age-related syndromes, such as frailty or sarcopenia, which are associated with an increased risk of help-dependency and morbidity (Umegaki, 2016).

The concept of ‘Aging well’ has been declared a global health priority by the World Health Organization, drawing attention to late-life health (Keevil and Romero-Ortuno, 2015). Identifying low-performing individuals with (even mild) motor impairments offers the opportunity to provide specific support and interventions, potentially preventing further decline in motor abilities (Talar et al., 2021). A longitudinal design would allow observing whether the low-performing participants are at specific risk for diseases/morbidity. The current data allows identifying ‘healthy’ individuals with reduced motor functions based on the (normative) values provided for a large cohort of participants (n = 444) with a broad age range (21–88 years; see Supplementary Table S2). Moreover, the current data facilitate the detection of deficiencies in patients with neurological (or other) diseases across different motor dimensions by using the documented stable ratios between the dominant and non-dominant hands or the ARAT.