Brainsight

A customized double-cone coil was attached to a Magstim Rapid (Magstim). We selected this device because it has been used in previous studies that stimulated the dACC^{27 (link),35 (link)}. For the location of the dACC, we chose the MNI coordinates (0, 36, 36), where we identified the maximum differential activity in the correlation with the inter-press time (Fig. 2d) between part learners and single learners. The coordinates were individually adjusted using the inverse normalization method embedded in SPM12 (normalization based on the inverse deformation field). We also individually localized the toe area in the primary motor cortex as the MNI coordinates (−8, −22, 64) following previous studies^{36 (link),37 (link)}. Brainsight (Rogue Research Inc.) was used to register the stimulus sites onto the individual T1 image. The Brainsight navigation procedure allowed for precise placement of the coil on the distal dACC and the reference toe area.

Subjects were comfortably seated in an adjustable armchair with a headrest. The left M1 target coordinates were individually localized in each subject based on hand knob location in T1-weighted image. This target position was marked on the subject’s anatomical MRI scan using Brainsight® (Rogue Research Inc., Montreal Quebec, Canada) software. The subject’s anatomical MRI scan was then co-registered with the subject’s head using frameless stereotaxy43 (link). The subject’s head position was assessed using the Polaris (Northern Digital, Waterloo, Canada) infra-red tracking system to measure the position of scalp landmarks (nasion, nose-tip, and intratragal notch of the left and right ears) visible on the subject’s MRI. The TMS coil was placed over the target brain area. The root mean square of difference between the co-registered anatomical landmarks estimated by the neuronavigation software was limited below 2 mm for each subject for improving accuracy. After anatomic co-registration, a figure-of-eight coil was placed tangentially to the scalp in an orientation inducing a posterior-anterior current perpendicular to the main course of the central sulcus. The individual coil positioning parameters were stored in the neuronavigation software.

All procedures matched those previously reported^{5 (link)}. Briefly, the scalp hotspot for the left first dorsal interosseous muscle (L. FDI) was first identified over the hand representation of the right motor cortex using single-pulse monophasic TMS (MagStim 200^{2 (link)}, MagStim Co. Ltd, UK) with a figure-of-eight coil held at ~45° relative to the mid-sagittal line, corresponding to a posterior-to-anterior current direction across the central sulcus^{42 (link)}. Resting motor threshold (RMT) was then determined using an automatic threshold-tracking algorithm (adaptive PEST procedure)⁴³ . After hotspot identification and thresholding, 600 single, monophasic open-loop TMS pulses were delivered to the scalp hotspot for the L. FDI at 120% of RMT (inter-pulse interval: 5 s with 15% jitter). This intensity was chosen to elicit a motor-evoked potential (MEP) on each trial, and on average was achieved with 54.35 ± 2.35% of maximum stimulator output. Single-pulse TMS was typically delivered in 6 blocks of 100 pulses. Due to the long stimulation and recording blocks, small breaks were often taken within each block upon subject request or TMS coil overheating. During TMS, EEG recordings were simultaneous obtained and coil position was monitored online using frameless neuronavigation (BrainSight, Rogue Research, Montreal).

Prior to stimulation, subjects were registered to their T1 image and target via fiducial points at the nasion and tragi. During stimulation, the relative position of the subject and coil were tracked in real-time using Brainsight (Rogue Research Inc., Montreal, Canada), a frameless stereotaxic neuronavigation system that uses reflective markers monitored with an infrared camera.

Structural magnetic resonance imaging (MRI) was performed in all participants for use during TMS neuronavigation. Participants received T1-weighted anatomical magnetic resonance imaging scan on a 3T scanner (GE Healthcare, Ltd., UK) using a 3D spoiled gradient echo sequence (Buss et al., 2020 (link)). T1-weighted anatomical MRIs were analyzed with Freesurfer 6.0 or 7.1 (documented and freely)¹ to obtain cortical gray matter thickness (GMT) measurements. One AD participant’s scan was excluded due to a Freesurfer segmentation error.
GMT measurements in all participants (n = 15 CU and n = 25 AD) was extracted for the left hemisphere mean cortical thickness (average left hemisphere GMT; LH Cortical Thickness) and left precentral thickness (from Desikan-Killiany atlas; Left Precentral Thickness) to serve as covariates in the subsequent nested regression analysis. In order to control for any differences in cortical atrophy, scalp-to-cortex distance (SCD) was measured using Brainsight™ (Rogue Research Inc., Canada) to calculate the in-plane distance between the motor cortex stimulation target and the participant’s scalp along the purported TMS trajectory on each individual’s T1-weighted MRI scan (Brem et al., 2020 (link)).