Wrist actigraphy in estimation of sleep and wake in intellectually disabled subjects with motor handicaps

We studied the applicability of wrist actigraphy to sleep-wake estimation in patients with motor handicaps.

Concomitant polysomnographic and actigraphic recordings (16-24 h) were compared in three groups: normally moving subjects with normal sleep (n=10), sleep-disordered subjects without motor handicaps (n=13) and sleep-disordered patients with different motor disabilities (n=16). The motor abilities of the subjects were determined by clinical evaluation using a grading scale from 0 to 10. Their actual daily activity was calculated from the recordings as average activity scores.

In the healthy subjects, the mean difference between actigraphic and polysomnographic total sleep estimation was negligible (-1 min), while in both sleep-disordered groups, sleep was highly overestimated by actigraphy. There was a significant correlation between the motor ability score and the discrepancy between actigraphy and polysomnography, but individual data points were highly scattered. A more consistent correlation was found between the average activity score/min in actigrams and the discrepancy of actigraphic with polysomnographic total sleep estimation (Spearman's r=-0.58, P=0.0001, n=39). When the recordings with very low average activity score were rejected from the analyses (two patients without and six with motor handicaps), the overestimation of sleep by actigraphy was reduced but it still remained in both sleep-disordered groups. The mean differences of total sleep between actigraphy and polysomnography were 72 and 121 min and the rank order correlation coefficients 0.80 and 0.71 in patients without and with motor handicaps, respectively. The median discrepancy in total sleep estimation was 6% in both sleep-disordered groups.

In subjects with rudimentary motor abilities, a standard actigraphy can produce a signal, which is related to the amount of sleep scored in polysomnograms. The sleep parameters obtained by the two methods are not equal, however. The inspection of actigrams is more reliable than the clinical scaling of motor abilities in predicting the applicability of wrist actigraphy.