Threshold-tracking SICF protocol fails validation against standard amplitude method

Threshold-tracking short-interval intracortical facilitation, in its currently published form, cannot replace conventional amplitude-based SICF as a measure of cortical excitability. That is the central conclusion of a head-to-head method comparison published in Clinical Neurophysiology (in press, available online 29 September 2026) by Hatice Tankisi of Aarhus University and an international group of co-authors including Hugh Bostock, Jean-Pascal Lefaucheur and collaborators in Canada, Chile, Denmark, France, Germany, Lithuania, Turkey and the United Kingdom.

Short-interval intracortical facilitation is among the paired-pulse transcranial magnetic stimulation measures of interest as a potential biomarker of cortical dysfunction in disorders such as Parkinson's disease and amyotrophic lateral sclerosis. The standard technique, A-SICF, quantifies the increase in motor evoked potential amplitude when a suprathreshold test stimulus — calibrated to evoke a 1 mV MEP at rest — is preceded by a subthreshold conditioning stimulus. The threshold-tracking alternative, T-SICF-1, measures instead the reduction in threshold for a 200 µV MEP when the conditioning stimulus is set at 95% of that threshold.

The investigators recorded A-SICF and T-SICF-1 twice from 60 healthy volunteers, delivering 10 stimulus pairs at each of 14 inter-stimulus intervals spanning 1 to 4.9 ms — the window in which I-wave summation produces facilitation. A subgroup of 20 subjects returned for testing with a newly designed protocol, T-SICF-2, in which stimulus intensities matched those of A-SICF in the absence of facilitation.

The results split cleanly. T-SICF-1 values were indeed less variable across repeated recordings than normalized A-SICF values. But variability was the only advantage. The correlation between T-SICF-1 and normalized A-SICF across the range of inter-stimulus intervals was poor (R = -0.586), and the threshold-tracking method registered appreciable facilitation at intervals where the amplitude method showed none. The two protocols, in other words, were not measuring the same physiological phenomenon.

The authors attribute the divergence to three design features of T-SICF-1: weaker test stimuli, stronger conditioning stimuli, and contamination by short-interval intracortical inhibition — the GABA-A-mediated mechanism that threshold tracking is ordinarily used to assay. Because of that contamination, they conclude that T-SICF-1, unlike A-SICF and T-SICF-2, is not a good indicator of I-wave facilitation.

The modified protocol fared better on validity. T-SICF-2 correlated strongly with normalized A-SICF across inter-stimulus intervals (R = -0.896), confirming that a threshold-tracking approach can index the same underlying facilitation as the amplitude method. The trade-off was precision: variability of T-SICF-2 was no better than that of the amplitude measurement it was designed to improve upon.

For laboratories running paired-pulse TMS panels, the practical message is straightforward. The published threshold-tracking SICF protocol should not be adopted as a substitute for amplitude-based SICF, despite its better repeatability, because its conditioning stimuli are too strong and its output reflects mechanisms other than I-wave facilitation. Groups that value the throughput advantages of threshold tracking — a single threshold estimate replaces batches of amplitude trials — will need to wait for a protocol that combines the validity of T-SICF-2 with the low variability of T-SICF-1. Whether such a protocol can be engineered, and whether it holds in patient cohorts with altered cortical excitability, remains the open question this study sets up.