Simultaneous EEG-fMRI studies are mainly based on standard GE-EPI sequences, due to its high speed and contrast sensitivity. However, alternative fMRI sequences with higher temporal resolution, using multi-slice⁴, or reduced acoustic noise, using minimal switching pulse sequences^1,2,3 can provide eminent advantages to this technique, improving the fMRI spatio-temporal resolution, the EEG signal’s quality by reducing the gradient induced artifact, and potentially the multimodal combination of EEG and BOLD signals.

In this work, three novel fMRI pulse sequences are evaluated together with simultaneous EEG acquisition, with the aim to reduce the induced EEG Gradient Artifacts (GA), the acoustic noise during scanning and to increase the fMRI spatio-temporal resolution. These sequences are: a) a modified GE-EPI with sinusoidal readout ¹, b) a T₂-prepared RUFIS (Rotating Ultrafast Imaging Sequence) ³ and c) a Multiband blipped CAIPI ⁴. Additionally, technical tests are performed in order to ensure the instrumentation and patient safety ⁵.

A spherical water phantom, with a 32-channels EEG BrainCap (Brainproducts, Munich, Germany), was imaged in a 3.0T MR750w system (GE Healthcare, Waukesha, WI), using a standard GE-EPI, followed by the three proposed sequences.

The acquisition parameters for the GE-EPI and GE-EPI with sinusoidal readout were as follows: FA = 90°; 3x3x3 mm full coverage with 40 slices, TR = 3000ms and TE = 30ms. For Multiband blipped-CAIPI only TR = 1500ms was modified (with multiband factor = 2). Finally, singleshot smoothed trajectory T₂-prep RUFIS was run with the following parameters: FA = 3º, 3x3x3 mm, TE_T2prep = 80ms, 380 spokes per segment/shot, defining a spiral trajectory in a 3D sphere and rotating 120º in each shot (TR_shot = 1s). Scan time for each sequence was 6 minutes.

Four different tests were performed in this study: 1) Specific Absorption Rate (SAR) of each sequence was estimated over 6 seconds and 6 minutes acquisition. 2) EEG electrode’s temperature was measured during scanning using MR compatible and calibrated equipment with two fluoroptic probes: one placed on Fz channel and the second monitoring the equilibrium temperature as reference. 3) Induced gradient EEG artifacts (GA) were analyzed off-line, comparing the artifact pattern and amplitude for every sequence using Brain Vision Analyzer 2.0. 4) Acoustic noise measurements were performed with a Bruel and Kjaer Type 4189 calibrated microphone placed inside the head coil. A-weighted average [LAeq dB(A)] and peak [Lpeak dB] sound pressure levels were measured during 20 seconds for each sequence and ambient noise.

1) Specific Absorption Rate: SAR measurements were similar for GE-EPI (avg=0,15 W/Kg; peak=0,28 W/Kg ) , modified GE-EPI (avg=0,15 W/Kg ; peak=0,29 W/Kg ) and multiband sequences (avg=0,17 W/Kg; peak=0,34 W/Kg). Slightly higher SAR values were observed for the T₂-prep RUFIS sequence (avg=0,26 W/Kg; peak=0,53 W/Kg), but still within safety limits.

2) EEG electrode’s temperature: Temperatures were considerably stable during all the sequences, without observing significant increments during the 6 min acquisition. The highest temperature values were found for the multiband blipped-CAIPI sequence (Tavg=15,22ºC ± 0,13ºC; T_peak=15,53ºC), but remained below the manufacturer recommended limits.

3) Induced gradient artifacts: Similar GA patterns were observed for the GE-EPI, modified GE-EPI and multiband sequences, while T₂-prep RUFIS showed a smoother pattern with significantly reduced amplitude (>10 times minor with respect to the GE-EPI induced GA). (Figure1)

4) Acoustic noise: Up to 121,2 dB were registered for the standard GE-EPI sequence, being close to the human pain level (≈120 dB). Measurements during the multiband sequence resulted in a similar outcome. The noise level was reduced in a mean of 6,1 dB(A) for the sinusoidal GE-EPI, due to its flatter gradient switching. Remarkable results were achieved for the T₂-prep RUFIS sequence, showing an overall noise reduction of 38,3 dB(A) with respect the GE-EPI, which corresponds to only 6,1 dB(A) over the ambient noise inside the scanner. All measurements are summarized in Figure 2.

This study guarantees the instrumentation and patient safety for simultaneous EEG-fMRI acquisitions using three novel pulse sequences: sinusoidal GE EPI, multiband blipped-CAIPI and single shot T₂-prep RUFIS.

T₂-prep RUFIS sequence revealed highly beneficial characteristics for the multimodal technique, not only for the reduced gradient artifact, which can significantly improve the EEG quality ⁶, but also for high temporal resolution, achieved by applying advanced reconstruction techniques for volume reconstruction per shot ², and low acoustic noise level presented. Auditory, resting-state and sleep studies can particularly benefit from this sequence. On the other hand, multiband fMRI sequences can achieve higher temporal resolution and high spatial resolution acquisitions with a whole brain coverage in ≈500ms, providing a promising solution for a better interrelation of EEG and BOLD signals.