Cortical hot spots and labyrinths: why cortical neuromodulation for episodic migraine with aura should be personalized

Abstract

Stimulation protocols for medical devices should be rationally designed. For episodic migraine with aura we outline model-based design strategies toward preventive and acute therapies using stereotactic cortical neuromodulation. To this end, we regard a localized spreading depression (SD) wave segment as a central element in migraine pathophysiology. To describe nucleation and propagation features of the SD wave segment, we define the new concepts of cortical hot spots and labyrinths, respectively. In particular, we firstly focus exclusively on curvature-induced dynamical properties by studying a generic reaction-diffusion model of SD on the folded cortical surface. This surface is described with increasing level of details, including finally personalized simulations using patient's magnetic resonance imaging (MRI) scanner readings. At this stage, the only relevant factor that can modulate nucleation and propagation paths is the Gaussian curvature, which has the advantage of being rather readily accessible by MRI. We conclude with discussing further anatomical factors, such as areal, laminar, and cellular heterogeneity, that in addition to and in relation to Gaussian curvature determine the generalized concept of cortical hot spots and labyrinths as target structures for neuromodulation. Our numerical simulations suggest that these target structures are like fingerprints, they are individual features of each migraine sufferer. The goal in the future will be to provide individualized neural tissue simulations. These simulations should predict the clinical data and therefore can also serve as a test bed for exploring stereotactic cortical neuromodulation.

Keywords: gyrification; migraine; neuromodulation; reaction-diffusion; spreading depression.

Figure 1

Empirical and postulated propagating visual field defects. (A,B) Two similar propagation patterns of self-reported visual field defects during migraine with aura; the attacks were 258 days apart. The red number indicate time in minutes starting from the first recognition of the migraine aura phase. (C) Kinematical description of SD wave pattern in the left primary visual cortex (V1). The V1 surface is assumed flat and described by the layout of visual field coordinates. (D) Right visual hemifield with the progressive wave pattern that matches the SD pattern in (C). Red number give time in minutes. (C,D) are modified from Dahlem and Müller (2003).

Figure 2

Principal folding pattern correlating with a cortical area defined by a functional field. Modified from study by Dahlem and Tusch (2012). (A) Medial side of the brain. One the occipital pole to the left, the calcarine sulcus is marked (think black line). (B,C) Coronal section through the occipital pole. Blue and red line traces the stria of Gennari as a marker of the primary visual cortex (V1), which is outlined in (A) by a dashed line. (D) Gross 3D shape of V1, about two-thirds of the surface area lie within the CS walls. The ventral and dorsal half of V1 are assumed to be symmetric with respect to the fundus, in particular, they extend equally long into the anterior direction, the shape of the walls are approximated with smooth profiles (see study by Dahlem and Tusch (2012) for details). (E) Light blue surface as in (D) with the retinotopic grid on top. (F) Visual field polar coordinates. (G) Gaussian curvature of V1.

Figure 3

Wave segments deflected by a bump. (A) Scheme with definitions of offset parameter d and deflection angle ϕ. (B) Deflection angle ϕ vs. offset d. (C–H) Numerical simulation of a positive deflection. (I–Q) Numerical simulation of head-on collision with wave destruction.

Figure 4

Simulation of spreading depression on an individually shaped cortical surface obtained by MRI from a migraine sufferer.

Figure 5

For a cortical surface ${\kappa }_{\text{geo}}\equiv \text{sign}(K)\sqrt{\left|K\right|}$ is shown on a color scale [− 1.56/mm, 1.56/mm], as indicated by the color bar. From right to left, the spot with largest negative Gaussian curvature K = − 2.343/mm is brought into focus by zooming closer twice, as indicated by the orange and yellow lines, respectively. It represents a potential nucleation point for SD.

Figure 6

(A) Working model ot the migraine-generator network with accessible intervention points (see text) (Dahlem, 2013). (B) Schematic representation of a cortical cross section with meninges and skull; not to scale. The effect of SD and its noxious signature could significantly vary with overall size of affected area and SD's gyral, sulcal, and laminar location. Future studies will be needed to design specific treatment stimulation protocols that take into account the noxious signature and its phase within the migraine cycle.