Mapping direct corticofugal motor and premotor projections with nTMS

Introduction:

The human motor system has for many years been subject to extensive studies. Its function and properties became a matter of intensive debate after Penfield first disclosed several motor areas with direct cortical stimulation (Penfield, 1952). Ablation studies in animals and lesion studies in humans then provided a better, yet incomplete, understanding of the function of the respective areas (Freund, 1985). Especially imaging studies have highlighted the importance of non-primary motor areas (NPMA) in motor learning and recovery from cortical or subcortical lesions of the motor system (Chen, Cohen Hallet, 2002). A recent study confirmed that navigated transcranial magnetic stimulation (nTMS) can locate direct corticofugal projections originating in NPMAs (Teitti, 2008). This offers the interesting perspective to directly investigate NPMAs neural circuitry under physiological as well as pathological conditions.

The aim of this study was to develope a reliable and accurate nTMS-mapping algorithm for NPMAs. We argue that a reliable algorithm should consider (i) remote electric field effects and volume conduction by nTMS, e.g. stimulating the NPMA does not induce distant effects in M1, (ii) motor evoked potential (MEP) latencies after NPMA stimulation should not be longer than after M1 stimulation, (iii) the maps should not be confounded by multiple local maxima in one representation and (iv) should differentiate primary and secondary motor representations. To test accuracy we compared the M1 hotspot with data from intraoperative stimulation and the NPMA results with results described in functional imaging.

Methods:

We mapped the motor system in 18 healthy subjects with nTMS. First, the precentral gyrus was mapped and the M1 first dorsal interosseus muscle (FDI) hotspot defined. A stimulator intensity to produce a 500µV mean MEP was subsequently used to map the middle and superior frontal gyrus. The remote electric field intensity over the M1 hotspot was kept at subthreshold levels and the highest response considered the NPMA hotspot. Electric field location, direction and intensity of applied stimuli and MEP responses during mapping were saved and subsequently analysed offline using custom software in the MATLAB environment. We applied Gaussian smoothing and Euclidean cluster-analysis to account for multiple local maxima and differentiate primary and second motor representations.

Results:

A NPMA motor cluster could be found anterior and medial to the M1 cluster in all subjects. The M1 and NPMA hotspot orientation differed significantly. Suprathreshold nTMS over the NPMA hotspot did not elicit MEP responses in M1at the respective estimated stimulation strength. MEP latencies did not differ for all recorded finger, forearm and arm muscles.

Clustering analysis of mapping data revealed either single or multiple local maxima.

Conclusion:

We have confirmed that identifying direct corticofugal responses by nTMS of NPMAs is feasible. The NPMA hotspot location suits data from imaging studies of the dorsal premotor cortex (dPMC) (Fink, 1997). The dPMC also possesses direct corticofugal projections (Dum, 1991; Freund, 1985). Corticocortical dPMC-M1 connections unlikely account for MEP responses from dPMC as latencies are not longer than those of direct M1 corticospinal projections.

Remote electric field effects do improbably confound the dPMC responses as their intensity exerted over M1 is subthreshold and the substantially different electric field orientation additionally weakens remote effects (Tranchina, 1986; Werhahn, 1994).

Corticofugal dPMC projections can be examined non-invasively with high temporal and spatial resolution by nTMS. Future studies should address their properties under physiological (e.g.motor learning) and pathological (e.g. motor recovery) conditions.