A novel model incorporating two variability sources for describing motor evoked potentials

Area of Science:

• Neuroscience
• Biophysics
• Computational Biology

Background:

• Motor evoked potentials (MEPs) are crucial for transcranial magnetic stimulation (TMS) assessments of corticospinal excitability.
• Input-output (IO) curves, derived from MEPs across stimulation strengths, characterize the corticospinal tract.
• Understanding MEP signal generation and variability is key for accurate physiological interpretation and statistical analysis.

Purpose of the Study:

• To introduce and evaluate a novel regression model for MEP IO curve analysis.
• To differentiate MEP variability into two independent sources acting before and after neural recruitment.
• To compare the novel model against traditional sigmoidal regression with a single variability source.

Main Methods:

• A novel dual variability source regression model was applied to measured MEP IO data from twelve subjects.
• The model incorporates a sigmoidal nonlinearity representing neural recruitment.
• Performance was compared to traditional sigmoidal regression models.

Main Results:

• MEP amplitude distributions varied significantly across stimulation strengths, violating assumptions of traditional models.
• The dual variability source model demonstrated superior fitting of IO curve characteristics.
• The new model effectively captured phenomena like changing distribution spread and skewness along the IO curve.

Conclusions:

• MEP variability is best explained by two distinct sources, likely related to initial excitation and later processes.
• The novel model provides more accurate and sensitive estimation of IO curve features, improving TMS applications.
• This approach extracts valuable physiological information about neural variability previously disregarded as noise and may extend to other brain stimulation techniques.