Table 1. Descriptive data.

	Descriptive data				Ljung-Box Test			Auto-correlation for the next n beats (b)	Dynamics
	Mean Inter-beat (Ib) Duration (in ms)	min Ib Duration (ms)	max Ib Duration (ms)	Total N° of beats	Value (a)	khi	Pvalue(at 1rst lag)		T value	R Square	P
Seq1	212.9	95	432	155	15	14.52	<0.01	2	4.17	0.09	<0.01
Seq2	253.9	86	489	34	na	0.09	0.76	na	−0.45	−0.02	0.66
Seq3	211.9	53	1156	34	na	0.12	0.72	na	0.69	−0.01	0.49
Seq4	222.7	130	312	47	15	16.5	<0.01	12	0.17	−0.02	0.87
Seq6	217.3	88	617	57	1	6.57	0.03	1	1.78	0.04	0.08
Seq7	267.4	79	470	71	na	0.37	0.54	na	0.1	0.0001	0.32
Seq8	274.6	96	1048	28	na	0.48	0.48	na	1.23	0.02	0.229
Seq9	258.2	66	709	71	15	9.8	<0.01	4	3.05	0.11	0.01
Seq10	276.3	96	421	32	na	0.002	0.96	na	−0.9	−0.01	0.37
Seq11	253.8	98	726	155	15	29.09	<0.01	10	2.37	0.03	0.02

Descriptive data on inter-beat (Ib) duration, number of beats, value and results of the Ljung-Box test, the auto-correlation test and the dynamic results for each sequence.

For example, in Sequence 1, the Ljung-Box test indicates a significantly non-random pattern at the first lag, checking for the next lags indicates that non-random patterns are detectable within up to 15 lags (Supplementary Fig. S2). The autocorrelation test indicates that whatever the position in the sequence, the time at which the next 2 beats will occur can be predicted (Supplementary Fig. S3). Here, the dynamic analysis shows a significantly positive linear trend, thus an increase in the duration between two beats (deceleration).

^(a)Number of lags within which non-random patterns can be detected (Supplementary Fig. S2).

^(b)Number of beats where dependency between two beats is statistically significant (Supplementary Fig. S3).