Effects of low-field magnetic stimulation on brain glucose metabolism

Abstract

Echo planar imaging (EPI), the gold standard technique for functional MRI (fMRI), is based on fast magnetic field gradient switching. These time-varying magnetic fields induce electric (E) fields in the brain that could influence neuronal activity; but this has not been tested. Here we assessed the effects of EPI on brain glucose metabolism (marker of brain function) using PET and 18F 2-fluoro-2-deoxy-D-glucose ((18)FDG). Fifteen healthy subjects were in a 4 T magnet during the (18)FDG uptake period twice: with (ON) and without (OFF) EPI gradients pulses along the z-axis (G(z): 23 mT/m; 250 mus rise-time; 920 Hz). The E-field from these EPI pulses is non-homogeneous, increasing linearly from the gradient's isocenter (radial and z directions), which allowed us to assess the correlation between local strength of the E-field and the regional metabolic differences between ON and OFF sessions. Metabolic images were normalized to metabolic activity in the plane positioned at the gradient's isocenter where E=0 for both ON and OFF conditions. Statistical parametric analyses used to identify regions that differed between ON versus OFF (p<0.05, corrected) showed that the relative metabolism was lower in areas at the poles of the brain (inferior occipital and frontal and superior parietal cortices) for ON than for OFF, which was also documented with individual region of interest analysis. Moreover the magnitude of the metabolic decrements was significantly correlated with the estimated strength of E (r=0.68, p<0.0001); the stronger the E-field the larger the decreases. However, we did not detect differences between ON versus OFF conditions on mood ratings nor on absolute whole brain metabolism. This data provides preliminary evidence that EPI sequences may affect neuronal activity and merits further investigation.

Figure 1

Diagram of the experimental procedures. Subjects were tested with ¹⁸FDG and PET twice (EPI stimulation and Sham). For each of these studies the subjects remained in a 4 T MRI instrument starting 15 min. prior and during the uptake period of ¹⁸FDG (25 min. after injection) for a total of 40 min. with their heads positioned such that the center of the EPI gradients was parallel to the axial plane that transected the upper part of the corpus callosum. For the “EPI stimulation” condition they were tested with EPI gradients ON and for the “Sham” condition they were tested with the EPI gradient OFF but with exposure to the recorded auditory noise from when EPI gradients were ON. At the end of the ¹⁸FDG uptake period subjects were taken out of the MRI and positioned in the PET scanner and imaging was started 35 minutes after ¹⁸FDG injection (10 min. after they were taken out of the MRI). Blood samples were collected 5 and 30 min. after ¹⁸FDG injection to quantify radiotracer in plasma and the behavioral assessments for mood states were made using the PANAS prior to and after completion of the EPI or the Sham stimulations.

Figure 2

Time-varying magnetic (B) and electric (E) fields produced by the Gz-EPI readout gradient.

Figure 3

Estimated strength of E-field in the brain generated by the EPI gradients. a. Diagram of the E-field induced by the EPI gradients pulses along the z-axis (G_z: 23 mT/m; 250 microsecond rise-time; 920 Hz), which are non-homogeneous and increase linearly from the gradient’s isocenter both in the radial (r) and the z directions. b. Sagital plane showing the distribution of the E-field and location of the axial planes (top, bottom and E = 0) shown in panel “c”. c. Axial planes showing the distribution of the E-field including the axial plane where E = 0. The E-field was strongest in the posterior and polar brain regions (top and bottom).

Figure 4

Statistical maps showing the SPM results for the comparisons of FDG images (Sham – EPI ON) for p < 0.05, corrected for multiple comparisons. Brain images were normalized to the axial plane where E = 0.

Figure 5

Brain metabolic changes as a function of the local E-field. Diagram showing the relationship between the changes in metabolism (% change) and the strength of the local E-field (V/m) for voxels where SPM showed significant differences between conditions (p < 0.05, corrected). Regression coefficient corresponded to r = 0.71, p < 0.0001; the stronger the E-field the larger the decrements in metabolism.