Multiple parietal pathways are associated with rTMS-induced hippocampal network enhancement and episodic memory changes

Abstract

Repetitive transcranial magnetic stimulation (rTMS) of the inferior parietal cortex (IPC) increases resting-state functional connectivity (rsFC) of the hippocampus with the precuneus and other posterior cortical areas and causes proportional improvement of episodic memory. The anatomical pathway(s) responsible for the propagation of these effects from the IPC is unknown and may not be direct. In order to assess the relative contributions of candidate pathways from the IPC to the MTL via the parahippocampal cortex and precuneus, to the effects of rTMS on rsFC and memory improvement, we used diffusion tensor imaging to measure the extent to which individual differences in fractional anisotropy (FA) in these pathways accounted for individual differences in response. FA in the IPC-parahippocampal pathway and several MTL pathways predicted changes in rsFC. FA in both parahippocampal and hippocampal pathways was related to changes in episodic, but not procedural, memory. These results implicate pathways to the MTL in the enhancing effect of parietal rTMS on hippocampal rsFC and memory.

Keywords: DWI; Episodic memory; Hippocampus; Resting-state functional connectivity; rTMS.

Figure 1.

A. Timing of procedures and stimulation sessions. Six of the participants received four sessions of stimulation. DW = diffusion-weighted, RS = resting-state. Time between sessions: ▲ = 0 – 36 hours, * = 24 hours, # = 7–14 days. B. Memory tasks. Episodic memory task: Participants encoded face-word pairs during the training block and were asked to recall the paired word when shown the faces at testing. Procedural memory task: Participants were shown 14 arrangements of four cards and learned the probabilistic relationship between the cards and two possible outcomes through feedback. During testing, participants were tested on their ability to produce the optimal response for each arrangement.

Figure 2.

ROIs and tracts for FA calculations. A. ROIs, including the precuneus (blue), IPC (green), retrosplenial cortex (cyan), parahippocampal cortex (red), entorhinal cortex (yellow), and hippocampus (orange) overlaid on a single participant’s brain. Tracts used to calculate FA are shown in purple and include the IPC-parahippocampus (B), IPC-precuneus (C), parahippocampus-entorhinal (D), entorhinal-hippocampus (E), IPC-retrosplenial (F), and IPC-hippocampal pathways (G).

Figure 3.

A. Distribution of stimulation points. Topographical distribution of stimulated voxels across the left IPC, shown on a template brain (TT_N27). Colors represent the proportion of voxels stimulated across participants. B. IPC group seeding approaches 1. Location of the six sampled seeds for 12 participants. 2. Region sampled to find the seed voxel maximally connected with the IPC for the remaining 12 IPC participants.

Figure 4.

Left: Scatter plots of FA-rsFC associations in tracts where the correlation values differed significantly by group. ROIs (right) include the IPC (green), parahippocampal cortex (red), entorhinal cortex (yellow), and hippocampus (orange) with corresponding tracts in purple. The lightning bolt represents stimulation location in the IPC. Arrows between images represent hypothetical path of signal from the IPC to the hippocampus.

Figure 5.

Correlations between FA and rTMS-induced changes in episodic and procedural memory for (A) the IPC-parahippocampus and (B) IPC-hippocampus pathways. Asterisks show significant correlation differences between groups.