Comparing brain activation associated with isolated upper and lower limb movement across corresponding joints

Abstract

It was shown recently that functional activation across brain motor areas during locomotion and foot movements are similar but differ substantially from activation related to upper extremity movement (Miyai [2001]: Neuroimage 14:1186-1192). The activation pattern may be a function of the behavioral context of the movement rather than of its mechanical properties. We compare motor system activation patterns associated with isolated single-joint movement of corresponding joints in arm and leg carried out in equal frequency and range. Eleven healthy volunteers underwent BOLD-weighted fMRI while performing repetitive elbow or knee extension/flexion. To relate elbow and knee activation to the well-described patterns of finger movement, serial finger-to-thumb opposition was assessed in addition. After identifying task-related voxels using statistical parametric mapping, activation was measured in five regions of interest (ROI; primary motor [M1] and somatosensory cortex [S1], premotor cortex, supplementary motor area [SMA] divided into preSMA and SMA-proper, and cerebellum). Differences in the degree of activation across ROIs were found between elbow and knee movement. SMA-proper activation was prominent for knee, but almost absent for elbow movement (P < 0.05); finger movement produced small but constant SMA-proper activation. Ipsilateral M1 activation was detected during knee and finger movement, but was absent for the elbow task (P < 0.05). Knee movement showed less lateralization in M1 and S1 than other tasks (P < 0.05). The data demonstrate that central motor structures contribute differently to isolated elbow and knee movement. Activation during knee movement shows similarities to gait-related activation patterns.

Figure 1

Pooled activation maps demonstrate the somatotopy of primary motor cortex (M1), cerebellum, and supplementary motor area (SMA). a,b: Contralateral M1 fields for fingers (cyan), elbow (red) and knee (yellow) are found in locations predicted by the M1‐homunculus. It is evident that upper extremity somatotopy is fairly symmetric between hemispheres. In contrast, there is asymmetry in knee representations, which may be an artifact of higher intersubject variability (size, location and strength) of the knee movement‐related contralateral M1 focus. c,d: In ipsilateral cerebellum, the knee representation is found anterior to elbow and hand representations. The latter two show substantial overlap (see Discussion). e,f: Somatotopic representations for hand (rostrodorsal), elbow (rostroventral), and knee (caudal) are observed in ipsi‐ and contralateral SMA‐proper (vertical line through the anterior commissure indicates the border between preSMA and SMA‐proper). Note that both images show the SMA‐proper of the same hemisphere while subjects move their left (e) or right limbs (f). Underlying anatomical images represent averages of T1‐scans of all subjects.

Figure 2

Activation maps of a representative subject performing left‐sided knee (yellow), elbow (red), and finger movement (cyan), demonstrate (a) premotor activation during finger movement. b: SMA activation during knee movement and M1 activation related to finger movement (white arrow indicates the position of the hand notch). c,e: Somatotopy in cerebellar anterior lobe with the knee representation anterior to elbow and finger areas. Additionally, finger‐related activation is found in the ipsilateral posterior lobe and contralateral anterior lobe of the cerebellum. d: Demonstrates somatotopy in the contralateral SMA‐proper (vertical line through the anterior commissure indicates the border between preSMA and SMA‐proper).

Figure 3

Mean activation indices (AI) plotted for each ROI and task. Knee movement activated SMA‐proper more than any other task. Contralateral M1 was strongly activated by finger movement, whereas knee movement produced little BOLD changes in contralateral M1. Ipsilateral M1 was active during knee and finger but not during elbow movement. PM was more activated by finger movement than by other tasks (*P < 0.05).

Figure 4

The lateralization index (LI) is plotted for different ROIs and movement tasks. FMRI‐activation during knee movement was less lateralized than activation associated with upper limb movement. This difference was significant for M1 and S1 (*P < 0.05). LIs for preSMA are omitted because less than half of the subjects showed activation in this area.