The overall purpose of this study was to establish an objective, repeatable, and quantitative platform with which to evaluate finger individuation that is translatable across multiple clinical and research settings. To this end, here we provide a set of normative values for three separate finger individuation scores in healthy adults using the Cyberglove III, each of which emphasize a different aspect of performance and subsequent optimal clinical applications.
While this data glove has been and continues to be used in many research settings, (1,2,7,25–28,31,32) this work enables more informed translational applications when objective measurement of finger individuation is needed.
In our analysis, we discovered that each score captures different aspects of individuation, and is therefore likely better applicable to certain clinical and research scenarios, as explored in the subsections below. To our knowledge, this is the first study demonstrating the repeatability and validity of techniques using the Cyberglove III to study kinematic finger individuation.
Thielbar Individuation Score
The cohort of healthy volunteers enrolled in our study predictably scored slightly higher than the cohort of patients with irreversible brain injury due to stroke in Theilbar and colleagues’ study (1) as shown in figure 4 of their results. While the Thielbar scores calculated in our study demonstrated overall repeatability, it varied by finger. For example, while the pinky scores were rather widely distributed, excellent repeatability was observed as the ICC was greater than 0.9. (33,34) There was ‘good’ repeatability of Thielbar individuation scores of the index, middle, and ring fingers as these ICC’s fit into the range of 0.75-0.9. There was ‘moderate’ repeatability of Thielbar individuation scores of the thumb as the ICC fell into the range of 0.5-0.75 (33). The low MDC’s will inform future studies aimed at determining the minimal clinically important difference (MCID) in patients with hand motor dysfunction, which will be critical to adapting these techniques to rehabilitative settings aimed at determining patients’ recovery trajectory.
Although moderately weak, the statistically significant relationship between Thielbar individuation scores and subjective ratings indicate that trials subjectively assessed as ‘good’ (1), ‘moderate’ (2), or ‘poor’ (3) followed a trend in Thielbar individuation score. Moreover, when the Thielbar individuation equation was used to calculate individuation scores with trials of participants closing their hand into a fist, scores were expectedly close to zero as all fingers were engaged in these trials. Ultimately, the Thielbar individuation equation offers the advantage of being suited to assess both healthy adults and patients with pathology of the motor system as it is sensitive enough to detect change in patients with stroke (1) and healthy adults predictably achieve scores higher than those with pathology. However, a notable disadvantage is the lack of emphasis on more nuanced components of motion, like range of motion, resulting in a diminished sensitivity to changes in individuals with minor dysfunction of the hand motor system.
Normalized Individuation Score
The normalized individuation equation was designed to promote and give credit for full range of motion, as individuation is most difficult at the end-range.(35,36) For the normalized individuation scores, there was good repeatability of the thumb, index, and pinky fingers and moderate repeatability of the middle and ring fingers. As participants were more able to reach their maximum range of motion across all trials, they achieved higher mean individuation scores. The rather low MDC’s indicate a high sensitivity to change in individuation ability that would be reflected in this individuation score.
There was a significant, moderate negative relationship between normalized individuation scores and subjective ratings, and the ranges of individuation scores within each subjective rating category clearly overlapped. Additionally, the normalized individuation equation resulted in uniformly low scores for trials of participants closing their hand into a fist, and the wider distribution was likely due to its greater weight on range of motion rather than merely the position of the fingers as utilized for the Thielbar equation. However, the predictably low scores for trials of closing the hand into a fist indicate this equation would result in low scores for poor performance of individuation in individuals with motor system pathology.
Compared to the Thielbar individuation scores, participants’ mean normalized scores were slightly higher. However, the distribution of these scores was much more dispersed between 0 and 1. This observed variability of normalized individuation scores is due to its emphasis on participants’ range of motion and how closely they reach their range of motion across all trials, making this equation especially suited to detect more subtle deficits in individuals with normal or near-normal hand function at baseline. Also, improvements or losses may be seen as range of motion changes through injury or rehabilitation. However, it should be noted that a wider range of normal-at-baseline scores must be taken into account. This may mean that it is the least optimal of the three equations to measure hand function in individuals with severe dysfunction of the motor system, such as stroke or proximal nerve damage, as uniformly low scores with poor range of motion render the examiner unable to detect slight changes in motor system performance, but it is the most optimal for detecting subtle changes and/or changes that encompass range of motion.
Threshold Individuation Score
For the threshold individuation equation, participants’ mean scores were the highest as this equation limits the impact of range of motion as a confounding factor, and instead assesses whether participants can perform the basic components of this movement. As these participants were all healthy, they typically could easily surpass this 50% threshold. There was good repeatability of the index, ring, and pinky individuation scores as well as moderate repeatability of thumb and middle fingers. The threshold MDC’s were also low, similar to the Thielbar and Normalized equations, and there is a need for larger studies of both healthy adults and patients with hand motor dysfunction.
In future studies and analyses, the threshold equation will likely prove useful in assessing patients with severe deficits in their motor system as scores for these individuals would be more widely distributed between zero and one. Therefore, this equation may capture slight changes in motor system performance as participants recover from a severe neurological injury, such as stroke. A weak but significant negative relationship was identified between individuation score and subjective rating. Additionally, the ranges of scores within each subjective rating clearly overlapped. Relationships between subjective ratings and individuation scores from all equations indicate the objective quantitative methods of assessing hand dexterity correspond to a subjective perception of performance quality. This is an important consideration as subjective methods have traditionally been used to assess dexterity in the clinical setting for many disturbances in the hand motor system. (3,8,13)
When threshold individuation scores were calculated from trials of hand open/close, scores were higher and more widespread compared to Thielbar and normalized scores for the same trials. Although participants’ individual trial scores were sometimes quite close to zero, their means were overall higher. So, while this equation has a unique utility in assessing dexterity in individuals with severe deficits, it is unable to produce scores that consistently reach zero when all fingers are engaged. Therefore, even in cases where dexterity is severely impaired, threshold scores will be higher compared to when the same data is used to calculate Thielbar and normalized scores. However, this may not be a true limitation, as the inability to reach zero may be due to the order in which the fingers were closed when forming a fist because the first finger to reach its threshold during trials of closing the hand into a fist was selected as the “indicated finger” for the purposes of calculating individuation scores.
Limitations
Our cohort of participants was limited to 20 rather young healthy adults, most of whom were female. Fortunately, it has been previously demonstrated that hand strength and function typically remain consistent until approximately 60 years of age,(37) so future renditions of this study with middle-aged adults would likely produce similar results. Notably, extrapolations to the pediatric and elderly population should be made with caution or not at all. Additionally, we did not study patients with known pathology such as osteoarthritis of the hand, peripheral nerve injuries, or nerve compressions, which can confound performance. (37,38) Therefore, extrapolation of these results to patients with musculoskeletal impairments should not be made. Although the techniques presented here were developed with the intention of assessing motor system dysfunction, the translation into assessing patients with dysfunction of any sort remains to be seen in future studies.
Second, although the results presented here depict statistical evidence for the repeatability of these methods, visual representations of scores show clear within-participant changes in individuation scores for the normalized and threshold individuation equations. It is important to note, however, that participants with large changes in individuation scores for a given finger often had similar changes for other fingers and for both the normalized and threshold equations. This suggests the changes in individuation scores across the two visits for a participant have multiple contributory factors, such as changes in effort, familiarity with the task, and slight changes in positioning both between participants and visits. This variability does not suggest poor performance of these scores or poor function the Cyberglove system in representing the individuation performance of a single trial, but instead suggests the needed acceptance of a larger range of “normal” values overall, and/or the need for repeat testing at each assessment to establish the individual’s true ability.