Transcranial magnetic stimulation (TMS) of the human frontal cortex: implications for repetitive TMS treatment of depression

Abstract

Repetitive transcranial magnetic stimulation (rTMS) is a noninvasive tool used to manipulate activity in specific neural circuits of the human brain. Clinical studies suggest that, in some patients with major depression, rTMS has the potential to alleviate symptoms that may be related to functional abnormalities in a frontocingulate circuit. This paper reviews the rationale for the use of rTMS in this context. The following topics are discussed: symptoms and cognition in major depression, with special emphasis on the initiation of speech; neuroimaging studies of depression; rTMS as treatment for depression; structure and function of the mid-dorsolateral frontal and anterior cingulate cortices; and combined TMS/positron emission tomography studies of frontocortical connectivity.

Fig. 1: Comparison of brain locations (in standardized stereotaxic space) identified as hypometabolic or hypoperfused in previous imaging studies of depression (labelled 2, 3 and 428) with those targeted in our transcranial magnetic stimulation (TMS)/positron emission tomography (PET) studies (labelled 129,30,31). The location of the mid-dorsolateral frontal cortex (MDLFC) is also indicated; in repetitive TMS (rTMS) studies of depression, this location is typically defined as a region located 5 cm anterior to the primary motor cortex (M1). Here, we established the “average” location of the MDLFC as follows: magnetic resonance images (MRIs) obtained in 152 healthy subjects were used to label the brain locations located 5 cm in front of the left M1 in each subject and to transform these “MDLFC” labels from native to standardized space. In the case of the anterior cingulate cortex (ACC), projections of brain locations onto a single sagittal slice were used.

Fig. 2: Schematic diagrams of the lateral (left), medial (middle) and inferior (right) surfaces of the human frontal lobe to illustrate its cytoarchitectonic parcellation. Reproduced with permission from Elsevier.

Fig. 3: Cytoarchitectonic subdivisions of human and monkey cingulate cortex. A: Along a rostrocaudal axis, the cingulate cortex can be divided into (1) a posterior region (areas 23, 26, 29, 30, 31) characterized mostly by a granular type of cortex and (2) an anterior agranular region (areas 24, 25, 32, 33). The anterior cingulate cortex is an agranular type of cortex (i.e., layer IV is absent) with a prominent and deeply stained layer V. Area 32 has an incipient granular layer IV. A ventrodorsal distinction, based on the degree of laminar differentiation, sets apart the old periallocortical areas adjacent to the corpus callosum (area 33) from the proisocortical region (areas 24, 25) and the paralimbic region on the upper bank of the cingulate sulcus and in the paracingulate gyrus (area 32). In addition to these main rostrocaudal and ventrodorsal distinctions, subtle variations in cytoarchitecture define further subdivisions of area 32, often reflecting structural features of the adjacent neocortical areas. The approximate position of the corticospinal fields is indicated relative to the vertical plane passing through the anterior commissure (VCA). VCP = vertical plane passing through the posterior commissure. B: Cytoarchitectonic areas superimposed on the flat map of the medial wall of the human brain. The bold lines outline the cingulate region, the thinner dashed lines show the borders between the cingulate areas (e.g., between areas 24 and 23), and the dotted lines indicate the borders between subdivisions of each area (e.g., between areas 24b and 24c). C: Location of the motor areas on the medial wall of the monkey brain. The dotted lines show the boundaries of the cytoarchitectonic areas. Shaded areas correspond to the territory of origin of corticospinal projections to cervical and upper thoracic segments. M1 = primary motor cortex, SMA = supplementary motor area, CMAr = rostral cingulate motor area, CMAd = caudal cingulate motor area, dorsal bank, CMAv = caudal cingulate motor area, ventral bank. Reproduced with permission from Macmillan Magazines Ltd. (www.nature.com/reviews) (Nat Rev Neurosci 2001;2:417-24).

Fig. 4: Modulation of corticocortical connectivity by rTMS. The flowchart (A) indicates the sequence of events during the TMS/PETstudies; the PET scans were repeated every 10 min (Base, no TMS applied; dpTMS, double-pulse TMS). The target site (B) within the left MDLFC was selected from a previous blood-flow activation study by Petrides et al; the “peak” is located just above the left inferior frontal sulcus. The proximity of the target site (crosshair) and the coil centre derived from the transmission scans in the 8 subjects (colour “bars”) demonstrate the successful positioning of the coil with frameless stereotaxy (C). After subtraction (D–F) and regression (G) analyses of blood-flow data, the images depict the exact locations of statistically significant decreases (D, E) and increases (F) in blood flow and significant positive correlation with blood flow at the stimulation site (G). The thresholded maps of t- statistic values (t > 3.0 or t < –3.0) are superimposed on coronal (D) and sagittal (E–G) sections through the average MRI of the 8 subjects. All images are aligned within the standardized stereotaxic space. Reproduced with permission from Blackwell Publishing (Eur J Neurosci 2001;14:1405-11).

Fig. 5: Dopamine release induced by rTMS of the prefrontal cortex. Left: organization of corticostriatal projections to the monkey putamen. Top right: changes in dopamine release and cerebral blood flow (CBF) in the human putamen after rTMS was applied over the left primary motor cortex (M1). Bottom right: changes in dopamine release in the human caudate nucleus after rTMS was applied over the left MDLFC. Location (red markers) of the 2 stimulation sites, the left MDLFC and the left occipital cortex, on the MRI of 1 subject in stereotaxic space. PMd = dorsal promoter, PMv = ventral promoter. Reproduced with permission from Springer-Verlag (Exp Brain Res 1998; 120:114-28).