Hand Dexterity and Pyramidal Dysfunction in Friedreich Ataxia, A Finger Tapping Study

ABSTRACT

Background

Loss of hand dexterity has a profound impact on disability in patients with cerebellar, pyramidal, or extrapyramidal diseases. Analysis of multiple finger tapping (FT) parameters can contribute to identify the underlying physiopathology, while providing a quantitative clinical assessment tool, particularly in patients not reliably evaluated using clinical rating scales. Here, we used an automated method of FT analysis in Friedreich ataxia (FRDA) to disentangle cerebellar (prominent FT rate variability), extrapyramidal (FT progressive amplitude reduction without slowing of tapping rate), and pyramidal (progressive decrease of FT rate and amplitude) contribution to upper limb loss of dexterity. FT parameters were then related to FRDA clinical parameters and upper limbs motor evoked potential (MEPs).

Methods

Twenty‐four FRDA patients and matched healthy subjects performed FT with the dominant hand for 90 seconds. FT rate, FT rate variability, FT amplitude, and linear regressions of FT movement parameters were automatically computed. Eleven patients underwent MEPs, measured at the first dorsal interosseous of the dominant hand to determine central motor conduction time (CMCT).

Results

FRDA patients had slower and more regular FT rate than controls. Eleven FRDA patients showed FT rate slowing. Those patients had longer disease duration and higher Scale for the Assessment and Rating of Ataxia (SARA) scores. Seven patients with FT rate slowing had MEP and all displayed prolonged CMCT, whereas the 4 other patients with constant FT rate had normal CMCT.

Conclusion

This study provides evidence for a prominent involvement of pyramidal dysfunction in upper limb dexterity loss as well as a potential outcome measure for clinical studies in FRDA.

Impaired performance of upper limb alternating movements was first described by Joseph Babinski as a cerebellar sign. ¹ Clinical experience showed that impaired performance of upper limb alternating movements also occurs in various central nervous system (CNS) disorders affecting the corticospinal tracts or the basal ganglia. ² , ³ , ⁴ In CNS disorders combining various degree of involvement of cerebellar, pyramidal, and extrapyramidal systems, determining that part of the CNS is mainly responsible for the patient's deficits may prove critical when considering symptomatic treatment, rehabilitation strategies, or the neural systems to target with restoring therapies, including gene therapy.

The localization yield of alternating movement impairment becomes substantially higher when its characteristics are objectively quantified. When using a finger tapping (FT) paradigm, this requires long periods of time (eg, 60 s) and a high number of movement cycles (eg, >50) performed at a comfort rate. ⁵ , ⁶ , ⁷ FT parameters include tapping rate, amplitude, and regularity, which may evolve over time as FT sequences are repeated. Whereas, a slow tapping rate is a shared feature of cerebellar, pyramidal, and both hyper‐ and hypokinetic extrapyramidal movement disorders, ⁶ , ⁷ , ⁸ , ⁹ other FT parameters are specific for different conditions. In early stage Parkinson disease (PD), tapping rate is slow but remains constant over time and it is accompanied by decreased FT movement amplitude. ⁵ , ¹⁰ , ¹¹ Conversely, pyramidal tract lesions cause a progressive decrease in both FT movement amplitude and tapping rate. ¹² In cerebellar disease, the most pronounced alteration is the loss of FT tapping regularity, with up to 5‐fold increase in the variance of the timing between consecutive tappings. ⁴ , ⁶ , ⁷

Friedreich ataxia (FRDA) is the commonest autosomal recessive ataxia in Caucasians. ¹³ The earliest FRDA neuropathology affects dorsal root ganglia (DRGs), posterior columns, and spinocerebellar tracts in the spinal cord, followed by progressive atrophy of the cerebellar dentate nuclei (DN) and efferent fibers, leading to a “tabeto‐cerebellar” ataxic pattern. ¹⁴ FRDA is also characterized by progressive atrophy of corticospinal tracts leading to weakness that becomes prominent in advanced disease. Almost always present at symptom onset, upper limb dysfunction is a major source of disability for FRDA patients ¹⁵ , ¹⁶ that keeps worsening even after patients lose the ability to walk. The clinical rating scales currently used to study the natural evolution of FRDA and because outcome measures in clinical trials (the Friedreich Ataxia Rating Scale [FARS] ¹⁷ and the Scale for the Assessment and Rating of Ataxia [SARA] ¹⁸ ) are much more accurate and sensitive in the assessment of motor function in ambulatory FRDA patients. In non‐ambulatory patients, both scales become noisier and less sensitive to disease progression, which at this stage is driven by upper limb function and speech impairment. Furthermore, advanced FRDA patients present with sensory, cerebellar, and pyramidal features that are difficult to disentangle clinically. ¹⁹

Appropriate quantification of upper limb dysfunction may therefore provide a robust measure of progression that has functional implications and remains valid throughout the entire course of the disease. In that framework, we developed a comprehensive, quantitative, and automated assessment of FT aiming at discriminating the role of different neural systems in causing loss of upper limb dexterity in FRDA. We used this paradigm to identify patterns of cerebellar (slow FT rate associated to prominent FT rate variability) versus hypokinetic extrapyramidal (slow but constant FT rate combined with FT progressive amplitude reduction) versus pyramidal (slow and progressively decreasing FT rate combined with progressive decrease of finger tapping amplitude) dysfunction and to correlate FT parameters with clinical, neurophysiological, and genetic parameters.

Methods

Participants

Twenty‐four FRDA patients from the Brussels site of the European Friedreich Ataxia Consortium for Translational Studies (EFACTS) clinical study ²⁰ , ²¹ and an equal number of healthy individuals matched for age, sex, and handedness (15 women; mean age = 25 years, range = 9–46 years; 3 left‐handed) without any prior history of neurological or psychiatric disease participated in the study. Main clinical characteristics are detailed in Table 1.

TABLE 1.

Characteristics of the included FRDA patients

Age (mean, [range], yr)	26 [9–46]
SARA (median, [range])/40	20 [4.5–32]
Upper‐limb items (median, [range])/12	6 [2–11]
Age of symptoms onset (mean, [range], yr)	12 [4–30]
Disease duration (median ± SD; yr)	13 ± 8.5
GAA1 (median, [range])	638 [280–1000]
Neurological examination
Abolition of upper‐limb deep tendon reflexes	24/24
Upper‐limb spasticity	0/24
Upper‐limb distal muscle atrophy	2/24
Mild	2/24
Upper‐limb distal weakness	2/24
Mild	2/24
Upper‐limb tactile perception impairment	0/24
Plantar responses, extension	20/24

GAA1, number of GAA1 triplet expansion on the shortest allele; SARA, score on the Scale for the Assessment and Rating of Ataxia; SD, standard deviation; upper‐limb items, finger‐nose, finger chase, and pronation/supination;.

Eleven FRDA patients (mean age = 31 years; range = 17–46 years; 5 women; mean SARA score = 21, range = 13–32; mean GAA1 triplet expansion = 621; range = 280–910; mean age of symptoms onset = 15 years; range = 7–30 years; mean disease duration = 17 ± 10 years) accepted to undergo motor evoked potentials (MEPs).

Experimental Paradigm

Participants performed repetitive right index finger–thumb oppositions at a self‐paced and comfortable rate for 90 seconds. At comfort rate (ie, self‐paced preferred rate, generally around 2 Hz), ⁵ healthy individuals and early stage PD patients do not show any significant decrease in FT rate over time. ⁵ , ¹⁰ , ¹¹ Recording over 90 s provides sufficient sensitivity to detect FT rate or amplitude changes over time. ⁶ , ⁷ To prevent potential biases because of subtle upper limbs proprioceptive impairment, the task was realized under visual control. Participants’ right index‐finger movements were monitored with a 3‐axis Cartesian reference frame attached to an accelerometer breakout board (Acc, ADXL335 iMEMS Accelerometer, Analog Devices, Norwood, MA) attached to the nail of their index finger with 3 M Medical Tape Micropore Paper 2″.

To assess potential pyramidal tracts involvement in FT alterations, central motor conduction time (CMCT) was evaluated using MEPs. MEPs were measured at first dorsal interosseous of the right hand and evoked with a 9 cm circular coil (Magstim 200 Co, Whitland, Dyfed, UK) placed around the central location (Cz), over the location that led to the best evoked motor responses. Motor threshold was assessed, and then 3–6 MEPs were recorded by stimulating at 140% of the motor threshold at the determined cortical location and at a cervical location above C5–C6. ²² MEPs were recorded within 2 weeks of the FT task and CMCT results were compared to established normal values. ²² , ²³ Abnormally long CMCT for upper‐limbs was defined as over 9.3 ms, corresponding to the mean normal values of the clinical neurophysiology department plus 2 standard deviations (SDs). ²² , ²³

Movement Analysis

Rate, Amplitude, and Variability Computation

The time‐course raw analogical values of the 3‐axis accelerometer data recorded during FT was processed by an in‐house Python program. The first step consisted of dimensional reduction, assuming that the motion occurs along a straight axis, the 3 accelerometers reading were combined into the so‐called reduced acceleration a(t), which is the projection of the acceleration vector onto the motion direction. A typical example of a(t) is shown in Figure 1, which shows motion cycles over 5 s for a typical healthy subject. One motion cycle occurs in the time interval comprised between blue dots, whereas the peaks indicated by green triangles correspond to the contact between index and thumb. The program automatically detects these events, using adaptive thresholds to account for subject‐induced irregularities in the accelerometer signal. From the peaks of the signal, the time interval between tapping is automatically extracted, and hence the rate. The amplitude motion of the finger is computed for each cycle by integrating twice a(t) between 2 cycle starting points. The amplitude for each cycle is then estimated by extracting the maximum of the integration result. To account for inter‐subject finger size variability, the estimated amplitude is normalized to its mean.

FIG 1.

Motion cycles over 5 s for a typical healthy subject. One motion cycle occurs in the time interval comprised between blue dots, whereas the peaks indicated by green triangles correspond to the contact between index and thumb.

Variability of FT rate and amplitude are obtained by computing their respective SD. To control for a rate effect on rate variability, a normalized rate variability was computed. Similarly, because movement amplitude reduction may mask rate slowing in pathological processes like early stage PD, a detrended rate variability was calculated. ⁸

Rate and Amplitude Trends

To determine whether a rate or amplitude was decreasing during the task, we performed standard linear regressions of the data as a function of time, yielding for each data types 4 parameters: the intercept, the slope, the Pearson correlation coefficient, and the P value of the fit. Only fits with a P value below 0.05 were included in the analysis. By definition, a given subject was then labelled as “decelerating” if the slope of the rate fit was negative and the Pearson correlation was larger than 30% in absolute value. Such 30% threshold was empirically set to get enough points for reliable correlations and to determine if the trend was sharp enough for clear deceleration. The same was done for significant amplitude reduction.

Amplitude Over Rate Effect

Because larger motions typically need more time to be accomplished, the rate variability was potentially dominated by the amplitude variability. To disentangle these effects, and therefore obtain a more intrinsic rate variability measure, we performed a linear regression of the rate with amplitude as predictor. The residuals of this model were defined as the reduced rates, whose SD was interpreted as an intrinsic rate variability measure (detrended rate variability).

Statistical and Correlation Analyses

Patients Versus Control Analysis

The normality of FT movement parameters was first assessed in each group using Q‐Q plots. FT parameters were compared between FRDA patients and healthy subjects. Levene test was used to seek for a violation of compared parameters equal variance assumption in the 2 groups. Welch test was used over Student t tests when the variances of the compared parameters were unequal. Adjusted P values for multiple comparisons were computed using the Bonferroni correction. After correction, significant P values were set at 0.008 for the comparisons between healthy subjects and FRDA patients.

Within FRDA Group Analysis

The same statistical method was used within the FRDA group to where FRDA patients were contrasted between those who displayed significant FT rate slowing or FT amplitude reduction to seek for potential differences in disease duration or SARA scores compared to FRDA patients with constant FT rate or amplitude. Adjusted P values for multiple comparisons were computed using the Bonferroni correction. After correction, significant P values were set at 0.017 for comparisons within FRDA group.

Data Availability Statement

De‐identified participant data will be shared as well as study protocol, statistical analysis, and analysis pipelines for data processing on reasonable request and after acceptance by institutional (ie, CUB Hôpital Erasme and Université libre de Bruxelles) authorities.

Results

FT Movement Parameters

Patients versus Healthy Subjects

Results are detailed in Table 2. Mainly, FRDA patients showed significantly slower FT rates and displayed significantly less FT rate and amplitude variability than healthy subjects. There was no difference in FT rate slopes nor FT amplitude slopes between the 2 groups. Fewer healthy participants showed FT rate slowing or amplitude decrease compared to FRDA patients and, when they did, changes were marginal compared to patients (4 times less rate slowing and 30 times less amplitude reduction).

TABLE 2.

Movement characteristics in FRDA patients and healthy participants

	FRDA patients	Healthy subjects	P
Rate (mean ± SD)	1.6 Hz ± 0.75	3.51 Hz ± 1.9	P < 0.001 a , b
Rate variability (mean ± SD)	0.22 ± 0.12	0.48 ± 0.44	P = 0.011 b
Rate variability detrended (mean ± SD)	0.24 ± 0.14	0.53 ± 0.44	P = 0.011 b
Rate slope (mean ± SD, °/ms)	−0.1265 ± 0.1	0.1027 ± 0.08	P = 0.435
Significant slowing (n%)	11 (46)	5 (21)	P = 0.005 a
Slowing slope (%/ms)	−0.12	−0.03	P = 0.02
Amplitude variability (mean ± SD)	0.6 ± 0.48	1.2 ± 1.17	P = 0.028 b
Amplitude slope (mean ± SD, %/ms)	−1.4 ± 4	0.1 ± 1	P = 0.13 b
Significant decrease (n%)	11 (46)	4 (17)	P = 0.003
Decrease slope (%/ms)	−0.03	−0.001	P = 0.05

SD, standard deviation.

^a Significant after correction for multiple comparisons, corrected P < 0.008.

^b Welch test was used for parameter comparison.

FRDA Intra‐Group Analysis

FRDA patients who showed FT amplitude decrease along FT tapping sequences (n = 11) had similar SARA scores compared to those whose amplitude was not decreased, both globally (19 ± 8 vs. 21 ± 8, P = 0.43) and in the items that assess upper limb function (6 ± 3 vs. 5 ± 2 p = 0.14).

FRDA patients who displayed FT rate slowing (n = 11) tended to have worse SARA scores (23 ± 7 vs. 17 ± 7, P = 0.033) and longer disease duration (16 years ± 11 vs. 10 years ± 4, P = 0.043), but there was no difference in SARA upper limb function items (6 ± 3 vs. 5 ± 2, P = 0.25). Of notice, a Welch test was used for disease duration comparisons.

CMCT Analysis

As a group, the FRDA patients who participated in MEPs investigations had prolonged CMCT (12.9 ± 4.4 ms; normal values <9.3 ms). ²² , ²³ However, CMCT was abnormally long (15 ± 3.5 ms; normal values <9.3 ms) only for those patients who displayed FT rate slowing (7/11), whereas those with no FT rate slowing had normal (2/11) or a slightly (9.6 ms ± 0.4) abnormal (2/11) CMCT compared to established normal values. ²² , ²³

Discussion

Prominent FT rate slowing without loss of FT regularity or decrease in amplitude suggests that pyramidal tract dysfunction is the main contributor of FT alteration in FRDA. This finding is likely to be generally valid despite the relatively small size of this study. Our sample consisted of patients from the Brussels site of the EFACTS clinical study, which currently includes >900 patients. No site effect was detected in the overall analysis of the EFACTS cohort ²⁰ , ²¹ and our sample is similar in terms of age, disease duration, SARA score, and size of GAA1 expansion to the average characteristics of the EFACTS cohort.

On average, FT rate was 50% lower in FRDA patients than in matched healthy individuals, in general agreement with previous studies in which FRDA patients were slower than healthy subjects in a number of tasks, such as the repeated pronation/supination item of the SARA, ¹⁷ “serious games,” ²⁴ and the items used to calculate the cerebellar composite functional severity (CCFS) score ¹⁹ (ie, a mathematical computation of the sum of pondered Z scores of the time to realize a 9HPT and a 10‐s clicks tasks). ²⁵ Like in this study, slower rate of alternation in the CCSF was associated to longer disease duration and higher SARA score. However, these previous investigations did not use methods that allow to identify the compromised neural system causing movement slowing. Even a semi‐automated score like the CCSF, which is now included in several ataxia studies, ²⁶ , ²⁷ has limitations in this regard because it focuses on the time needed to perform brief manual tasks, a parameter that alone cannot discriminate among different causes of movement slowing. ⁶ , ⁷ , ⁸ , ¹² , ²⁸

We used a method that analyzes multiple FT movement parameters, allowing to delve deeper into the pathophysiology underlying the observed movement slowing in FRDA patients. First, instead of showing the prominent FT rate variability that characterizes FT impairment in cerebellar disease, ⁴ , ⁶ , ⁷ FRDA patients were even more regular than healthy individuals. This finding cannot be accounted for either by lower FT rate; because rate normalized variability was similar to healthy participants, nor by FT movement amplitude reduction, which was specifically controlled for. These findings indicate that either cerebellar impairment may not be the sole cause of FT impairment in FRDA or that this slow FT rate reflects a compensatory strategy to achieve accuracy despite cerebellar dysfunction.

FRDA patients as a group also did not show the steep slope of amplitude reduction that characterizes early stages PD. ⁸ , ¹⁰ , ²⁹ Even though we cannot formally exclude that FRDA patients may have basal ganglia dysfunction similar to advanced PD patients who do not show such progressive FT amplitude decrease, this is unlikely, because neither FRDA patients in general nor FRDA patients in our cohort displayed prominent extrapyramidal features. Furthermore, though basal ganglia abnormalities, including volume loss and increased iron content, are described both in structural ³⁰ and functional ³¹ brain MRI studies performed in FRDA patients, they are much less prominent than cerebellar and pyramidal abnormalities.

FRDA patients consistently showed significant FT rate slowing along the tapping sequence, which is typical of pyramidal impairment. ¹² This abnormality was more severe in those with higher SARA scores and disease duration. Progressive pyramidal tracts involvement in FRDA is supported by neuropathological, clinical, and electrophysiological findings ³² , ³³ Weakness, spasticity, and muscle atrophy of the hands are eventually found in over two‐third of cases. ³⁴ Accordingly, we found that FT rate slowing was more severe in patients with higher global SARA scores, reflecting more advanced disease. FRDA patients with longer disease duration also showed more severe pyramidal tract dysfunction as assessed by MEPs, with more marked increase of CMCT. ³⁵ , ³⁶ Accordingly, in our study, patients who displayed significant FT rate slowing also had longer CMCT. The hypothesis that pyramidal tract pathology is the main contributor to FT abnormalities in FRDA is also supported by functional imaging studies highlighting decreased activation in the primary motor cortex in addition to the cerebellum in FT tasks, ³⁷ associated to compensatory increased activation in premotor cortical areas in FRDA patients at the lower end of clinical severity. ³⁸

Spinal cord posterior column atrophy is characteristic of FRDA neuropathology and upper‐limb proprioceptive impairment could also have played a role in FT rate alterations. ³⁹ However, the FT task was carried out under visual control, minimizing the possible contribution of proprioceptive loss. Furthermore, proprioceptive impairment in FRDA occurs early, is mostly stable over time and does not correlate with SARA scores nor with disease duration. ⁴⁰ , ⁴¹

On a methodological standpoint, the dedicated paradigm, the device, and the automated pipeline developed in this study provide an original and user‐friendly tool for the assessment of patients with movement abnormalities, contributing to discriminate among cerebellar, pyramidal, and extrapyramidal pathology. After this proof‐of‐principle study in FRDA patients, the exact yield, specificity and sensitivity of the method needs to be established by additional studies on patients with different underlying pathologies.

In FRDA, this approach promises to provide a quantitative assessment tool that is sensitive to disease progression throughout the course of the disease, including in non‐ambulatory patients who are not reliably assessed with the FARS or the SARA. However, given our limited sample size, validation via further studies involving larger patient cohorts is needed.

Author Roles

(1) Research project: A. Conception, B. Organization, C. Execution; (2) Statistical Analysis: A. Design, B. Execution, C. Review and Critique; (3) Manuscript: A. Writing of the First Draft, B. Review and Critique.

G.N.: 1A, 1B, 1C; 2B; 3A, 3B.

A.R.: 1B, 1C; 2A, 2C, 3B.

M.P.: 2C, 3B.

X.D.T.: 2C, 3B.

Disclosure

Ethical Compliance Statement: All participants were included in the study after written informed consent. The study had prior approval by the CUB Hôpital Erasme Ethics Committee and was performed in accordance with the Declaration of Helsinki.We confirm that we have read the Journal's position on issues involved in ethical publication and affirm that this work is consistent with those guidelines.

Funding Sources and Conflict of Interest: G.N. and X.D.T. are Postdoctorate Clinical Master Specialists at the Fonds de la Recherche Scientifique (FRS‐FNRS, Brussels, Belgium). Massimo Pandolfo received research grants from the Friedreich Ataxia Research Alliance (FARA). G.N., A.R., M.P., and X.D.T. report no biomedical financial interests or potential conflicts of interest relating to this work.

Financial Disclosures for the previous 12 months: No authors have received any funding from any institution, including personal relationships, interests, grants, employment, affiliations, patents, inventions, honoraria, consultancies, royalties, stock options/ownership, or expert testimony for the last 12 months.