Uncertainty quantification in transcranial magnetic stimulation via high-dimensional model representation.

Area of Science:

• Computational neuroscience
• Biomedical engineering
• Medical physics

Background:

• Transcranial magnetic stimulation (TMS) is a non-invasive brain stimulation technique.
• Accurate modeling of electric (E) fields induced by TMS is crucial for therapeutic efficacy.
• Quantifying uncertainties in TMS parameters is essential for reliable treatment outcomes.

Purpose of the Study:

• To develop and validate a computational framework for uncertainty quantification in TMS.
• To identify critical sources of uncertainty affecting TMS-induced E-fields.
• To provide insights for optimizing TMS application in clinical settings, such as depression therapy.

Main Methods:

• Leveraged high-dimensional model representations (HDMR) to approximate E-fields.
• Employed a multielement probabilistic collocation (ME-PC) method for component function approximation.
• Utilized a quasi-static finite-difference simulator for computing E-fields at collocation points.

Main Results:

• The proposed framework requires significantly fewer simulations than Monte Carlo methods.
• Uncertainties in tissue conductivities minimally impact TMS operation.
• Variations in coil position/orientation and brain size significantly affect induced E-fields.

Conclusions:

• Coil positioning and orientation uncertainty may decrease patient response rates in TMS therapy.
• Targeting the crest of a gyrus maximizes stimulation.
• Increased scalp-to-cortex distance reduces the magnitude of E-fields in the cortex.