Multifractality of human upper-limb velocity tracks motor optimization via hand dominance experiment

Abstract

Theoretical goal-directed movements follow a unique, stereotyped velocity profile, but continuous movement has no unique optimal solution, so it is unclear whether motor optimization makes continuous movement more stereotyped or more diverse. Previous studies indicated higher information content in the dominant hand, suggesting greater diversity. Because continuous velocity is a composition of overlapping joints movement of varying form, timing and amplitude, comparable to the events that compose spike trains and heartbeat series, we asked whether this diversity appears as a wider multifractal spectrum of velocity. We applied multifractal detrended fluctuation analysis to motion-capture velocity of cursive handwriting from right- and left-handed participants writing with both hands; hand dominance is a natural experiment in which task and geometry are fixed while optimization varies. With the dominant hand, spectra were wider, had a lower dominant exponent and were more positively asymmetric in both groups, whereas hands with the same functional role did not differ between groups. A logistic regression on the three spectral features identified the dominant hand with a cross-validated AUC of up to 0.93. Optimized continuous movement is thus not more stereotyped but more diverse: complex, neither stereotyped nor random, with multifractality that tracks hand dominance.