Figure S1. Tapping tempo and variance. Legend conventions as Fig. 2. (a) Mean tapping tempo across sessions. (b) Tapping variance remained consistent between forced and free sessions across three tapping blocks. However, natural tapping exhibited the smallest variance during the transition stage, ranging from 60ms to 70ms, marked by a sudden shift from constrained to free tapping.
Figure S2. Evaluating the consistency and reliability of natural tapping tempo. Legend conventions as Fig. 2. (a-c) Consistency between periods over 3 months with 7-day, 30-day, and 3- month intervals after the first participation. (a) Averaged effects among 32 participants. (b) Averaged effect and individual data. (c) Each connected line represents one participant with most of them showing consistent tempo preference over the 3 months, while others may have pronounced fluctuations in tempo. (d-g) Consistency between sessions. Natural tapping tempo remained consistent between two sequential sessions in four continuous entries over 3 months with 7-day, 30-day, and 3-month intervals after the first participation. (h-j) Reliability across states was observed. High correlations between time points were identified in both sessions (h-i) and when combined and averaged across two sessions (j).
Figure S3. Correlation between ITPF/IAPF with tapping tempo. Legend conventions as in Fig. 3 and Fig. 4. (a-c) Individual delta/theta peak frequency from the motor cortex during the resting state shows no correlation with tapping tempo. (d-f) Individual alpha peak frequency during resting state only shows correlation with natural/free tapping tempo (e) but not fast or slow tempo in both forced and free sessions (d, f).
Figure S4. ERP results and correlation with tapping tempo. Legend conventions as in Fig. 5. (a) Natural tapping showed a uniquely enhanced amplitude from 83-125ms. (b) ERP amplitudes were not correlated with fast or fast-to-free tapping tempo across participants. (c) ERP amplitudes were not correlated with slow or slow-to-free tapping tempo across participants.
Figure S5. (a) Time-averaged spectrum. Legend conventions as in Fig. 6. Natural tapping induced stronger power at the theta (b) and alpha (c) bands compared to the constrained tapping blocks.
Figure S6. Correlation between tapping-evoked IAPF and tapping tempo. Legend conventions as in Fig. 4. Tapping-evoked IAPF had a unique correlation with the natural tapping tempo (b) compared to constrained tapping in both forced and free sessions (a, c).
Figure S7. Source activity in CEN and salience network. Legend conventions as in Fig. 7. (a) Source activity in the CEN was significantly decreased during natural tapping (a) over a time course of 140-170ms (b). Source activity in the salience network did not show any changes between sessions or across tapping blocks (c, d).