Article Figures & Data
Figures
- Figure 1.
Experiment setup, conditions, timing, and trial-related brain activity. A, Setup from side view. The participant's head is tilted to permit direct viewing of objects on the platform. B, Experimental apparatus and graspable object shown from the participant's point of view. The same object (consisting of a smaller cube attached atop a larger cube) was always presented at the same location on the platform for every trial. Green star with dark shadow represents the fixation LED and its location in depth. Hand is positioned at its starting location. Right, The three different hand movements. C, Timing of one event-related trial. Trials began with the 3D graspable object being illuminated while the subject maintained fixation (preview phase; 6 s). Subjects were then instructed via headphones to perform one of three hand actions: grasp the top cube (Top), grasp the bottom cube (Bottom), or touch both cubes with their knuckles (Touch). This cue initiated the plan phase portion of the trial, in which, in addition to having visual information from the object, subjects also knew which hand action they were to perform. After a delay interval (10 s), subjects were cued (via an auditory beep) to perform the instructed hand movement (execute phase). Two seconds after the movement, vision of the object was extinguished and participants waited for commencement of the following trial (14 s, ITI). D, Example event-related BOLD activity from parietal cortex (posterior IPS) over the length of a trial. Events in D are time locked to correspond to events in C. Pattern classification analysis was performed on single trials based on the windowed average of the percentage signal change (% SC) corresponding to the three different time points denoted by each of the gray shaded bars (each corresponding to activity elicited from the 3 distinct trial phases preview, plan, and execute).
- Figure 2.
Decoding of object-directed movement intentions across the parieto-frontal network. Cortical areas that exhibited larger responses during movement planning than the preceding visual phase (plan > preview) are shown in orange/yellow activation. Results calculated across all subjects (random-effects GLM) are displayed on one representative subject's inflated hemispheres. The general locations of the selected ROIs are outlined in circles (actual ROIs were anatomically defined separately in each subject). Each ROI is color coded according to the pairwise discriminations they can decode during the plan phase (found in Fig. 4); see color legend at top for classification profiles (colors denote significant decoding accuracies for upcoming actions with respect to 50% chance). Sulcal landmarks are denoted by white lines (stylized according to the corresponding legends below each brain). LH, Left hemisphere; RH, right hemisphere.
- Figure 3.
Trial-related percentage signal change fMRI activations from each of the 14 plan-network ROIs and three sensory control ROIs. Activity in each plot is averaged across voxels within each ROI and across subjects. Plots show the profiles of typical preparatory activity found throughout parieto-frontal network areas. Vertical dashed lines correspond to the onset of the preview, plan, and execute phases of each trial (from left to right). Shaded gray bars highlight the two-volume (4 s) windows that were averaged and extracted for pattern classification analysis. Note that time corresponds to imaging volumes (TR of 2) and not seconds.
- Figure 4.
Classifier decoding accuracies for each ROI for the three trial phases (preview, plan, and execute; the middle 3 bars correspond to accuracies elicited during the plan phase). Error bars represent SEM across subjects. Solid black lines are chance accuracy level (50%). Black asterisks assess statistical significance with two-tailed t tests across subjects with respect to 50%. Note that no above-chance decoding is shown during the preview phase, when subjects were unaware which movement they were going to perform. Red asterisks denote statistical significance with paired two-tailed t tests for decoding accuracies across subjects for plan and execute phases with respect to within-trial decoding accuracies found during the preceding preview phase (i.e., assessing where accuracies are higher than that for simple object visual presentation, when subjects were unaware which action they would be performing). Importantly, any areas showing significant decoding during the plan phase with respect to 50% also show significant decoding with respect to the permutation tests (see Materials and Methods) and preview phase. Note that accurate classification can only be attributed to the spatial response patterns of different planned movement types and not the overall signal amplitudes within each ROI (see Fig. 5). Also note that decoding accuracies are color coded according to pairwise discriminations and not trial types.
- Figure 5.
No fMRI signal amplitude differences found within the parieto-frontal regions used for pattern classification. Responses are averaged across voxels within each ROI and across subjects (2-volume averaged windows corresponding to preview, plan, and execute phases). Note that only one significant univariate difference is observed in R-SPOC, a non-decoding region. Error bars represent SEM across subjects.
- Figure 6.
Classifier decoding accuracies in non-brain control regions. A, Non-brain control ROIs defined in each subject (denoted in green; example subject shown). B, Classifier accuracies for the three trial phases for the right ventricle (left) and outside the brain ROI (right). Error bars represent SEM across subjects. Solid lines show chance classification accuracy (50%). Importantly, no significant differences were found with t tests across subjects with respect to 50% chance.
Tables
- Table 1.
ROIs with corresponding Talairach coordinates (x, y, and z center of mass mean and SD)
ROI name Talairach coordinates ROI size Mean x Mean y Mean z SD x SD y SD z mm3 n voxels R-SPOC 4 −70 38 4.6 3.5 4.1 1782 66 L-SPOC −6 −74 36 4 3.9 3.6 2189 81 L-aPCu −14 −74 44 3.7 3.8 3.4 1895 70 L-pIPS −16 −63 50 3.3 4 3 1996 74 L-midIPS −35 −57 42 3.6 4 4.3 2053 76 L-post aIPS −36 −44 46 3.9 3.9 3.8 2094 78 L-aIPS −49 −34 44 3.8 4 4 1926 71 L-SMG −58 −41 29 3.6 3.3 4.2 1782 66 L-motor cortex −33 −19 56 4.5 4.4 3.6 1278 47 L-PMd cortex −23 −9 58 4.1 4 3.6 1914 71 L-precentral gyrus −39 −11 55 3.5 3 3.2 1679 62 L-PMv cortex −53 4 31 3.1 2.7 2.9 1617 60 L-preSMA −8 4 41 3.2 4 4 1896 70 L-SMA −7 −3 51 3.5 3.9 3.4 2026 75 L-SS cortex −41 −32 54 3.8 3.9 4.2 1528 57 L-HG −57 −26 6 4 3.7 4.1 1956 72 R-HG 57 −21 7 3.9 3.6 3.2 1627 60 -
Mean ROI sizes across subjects from ACPC data (in cubic millimeters and functional voxels).
-
In this issue
