Details

Autor(en) / Beteiligte

• Julkunen, P
• Kallioniemi, E
• Säisänen, L
• Könönen, M

Titel

P 207. Reliability of an automated protocol versus manual interpreters in analysing cortical silent period

Ist Teil von

• Clinical neurophysiology, 2013-10, Vol.124 (10), p.e162-e162

Ort / Verlag

Elsevier Ireland Ltd

Erscheinungsjahr

2013

Quelle

Alma/SFX Local Collection

Beschreibungen/Notizen

• Introduction Cortical silent period (cSP) is measured after shortly interrupting active muscle contraction with transcranial magnetic stimulation (TMS) (Fuhr, 1991). The cSP is a measure of cortical inhibition and representing interneuron inhibitory effect at excited motor cortical areas. Several pathological conditions and pharmacological manipulations induce changes to cSP duration. In addition, cSP has exhibited prognostic value e.g., during stroke recovery (Curra, 2002). It has been suggested that input–output characteristics of cSP be determined for thorough assessment of inhibitory interneurons (Werhahn, 2007; Kimiskidis, 2005). These characteristics are commonly analyzed manually from measured electromyography (EMG) signal. However, to avoid inter-interpreter effects in cSP interpretation and detection, as well as to allow quick measurement in real-time, an automatic analysis routine would be preferable. Methods We reanalyzed previously manually analyzed cSPs (Säisänen, 2008) of the right hand of 55 healthy subjects (27 male, 28 female, age range: 23–80) using a novel automatic routine. Five cSPs were induced at 120% of the resting motor threshold (rMT) focused on the left M1. Furthermore, we recruited one female (age 28) subject for whom the cSPs were induced with several stimulation intensities (SIs), and those cSPs were analyzed manually by two of the authors as well as using the automatic routine. In the automatic routine, we computed the first time-derivative of the single-trial EMG signal (smoothed), and detected longest interval with lower than 15 μV difference between consecutive samples. Results We found that 99% of all cSPs were identified correctly by the automated routine. There was a good agreement between the cSP durations analyzed manually and automatically (ICC = 0.972, p < 0.001, Fig. 1 ). Based on ANOVA, the automatic and manual routine did not exhibit significant differences, while the between-subject effect was significant ( p < 0.001). When studying the effects of SI and analysis type, we found SI had an effect on the cSP duration ( p < 0.001), but not the analysis type (manual or automatic). Between two interpreters, the agreement in manually analyzed cSP durations was excellent (ICC = 0.990, p < 0.001, 95%CI for the difference: 12 ms), as it was between the automatic and manual analysis (mean of two interpreters) (ICC = 0.990, p < 0.001, 95%CI for the difference: 12 ms). The difference between interpreters was similar to that between manual and automated analysis ( Fig. 1 ). Conclusions Use of automatic cSP detection may enable new mapping modality based on cSP duration, e.g., in patients with lowered cortical excitability, the cSP having a lower threshold than motor evoked potential (Werhahn, 2007). Also, automatic analysis routine will allow for a quick assessment of input–output curve for the cSP improving its applicability.