Introduction

Echo planar imaging (EPI) has been the gold standard for functional MRI (fMRI) for three decades. EPI-based fMRI approaches, however, have shortcomings, such as high acoustic noise and sensitivity to susceptibility artefacts and subject motion. To overcome these limitations, zero echo time fMRI modalities have been introduced, including the novel method Multi-Band Sweep Imaging with Fourier Transformation (MB-SWIFT)^1,2. MB-SWIFT has high excitation and acquisition bandwidths, practically zero echo time, it is quiet, insensitive to image distortions and motion artefacts, and produces a functional contrast comparable to EPI. Accordingly, this approach enables study designs which are extremely challenging with EPI, such as deep brain stimulation fMRI¹, spinal cord fMRI³, and behavioral studies in head-fixed rats⁴.
However, fundamental questions need to be addressed regarding the use of MB-SWIFT for functional imaging. At zero echo time, BOLD contrast disappears, so other functional contrast mechanisms need to be considered. Therefore, although the relationship between the BOLD fMRI signal and electrical brain activity has been studied extensively⁵, the findings of these studies can not necessarily be extrapolated to MB-SWIFT fMRI. The aim of this work was to study the correlation of electrical brain activity with the MB-SWIFT fMRI signal. For this purpose, we conducted simultaneous electrocorticography (ECoG)-fMRI studies in anesthetized and awake rats during electrical or mechanical whisker stimulation at a wide range of stimulation frequencies.

Methods

Animal procedures were approved by the Animal Experiment Board in Finland. Under isoflurane anesthesia, 7 male and 4 female Sprague-Dawley rats (n=11) underwent surgery for the placement of a head-fixation implant^2,4, four intracranial tungsten wire electrodes, and 4–5 skull screws over the cerebellum and olfactory bulb to serve as reference/ground electrodes and anchors for the implant (Fig. 1A, 1B). After 1–3 weeks of recovery time, the rats were habituated to awake fMRI^4,6.
fMRI was performed as previously described⁴. fMRI data were acquired at 9.4T Agilent MRI system with transceiver surface RF coil using 3D MB-SWIFT (2047 spokes, TR 0.97 ms, temporal resolution 2 s, 192/384 kHz excitation/acquisition bandwidths, 5-6° flip angle, and 64³ matrix size with 625 µm isotropic resolution). Simultaneous electrical recordings were performed with an MRI-compatible BrainAmp system. Data were processed and analyzed as described earlier^2,4. Each rat underwent 2–5 sessions of imaging in either the awake state or under isoflurane (0.5%) + medetomidine (0.03 mg/kg/h, sc) anesthesia.
The stimulation paradigm consisted of 16-s blocks of unilateral whisker stimulation at frequencies of 1, 3, 5, 7, 9, 13, and 17 Hz. Four blocks were delivered per frequency, with all blocks randomized by order and interposed with 44-s breaks. In separate experiments, both mechanical and electrical stimulation was used. Mechanical whisker stimulation was performed with puffs of air directed at a piece of adhesive tape on the whiskers (Fig. 1B). Electrical stimuli were delivered via two subcutaneous needles placed in the whisker pad between rows A–B and C–D (Fig. 1B, 1C). In awake rats, only mechanical whisker stimulation was used, and frequencies 3 and 7 were left out to shorten the imaging time. This resulted in three separate experimental conditions: electrical whisker pad stimulation in anesthetized rats (ISO+MED, electrical), and air puff-mediated whisker stimulation in anesthetized (ISO+MED, air puff) and awake rats (Awake, air puff).

Results

In anesthetized rats, the sum of somatosensory evoked potentials (SEP) and the area under the curve (AUC) of the fMRI response showed similar dependence on stimulation frequency, with response size typically peaking at 5–7 Hz and decreasing at higher frequencies. (Fig. 2, Fig. 3B, 3C). MB-SWIFT fMRI accurately illustrated the whole whisker-to-cortex sensory pathway, highlighting the large 3D-coverage and robustness of the approach (Fig. 3A). The relationship between electrical recordings and the MB-SWIFT fMRI signal under anesthesia was well explained by a linear model (Fig. 4). In the awake condition, the sum of SEPs typically peaked at 9 Hz, accompanied by a less dramatic drop for the higher frequencies than in anesthetized rats (Fig. 2B). Although the awake fMRI response seems to increase in the lower frequency range akin to the sum of SEPs, a clear deviation is seen at 13 and 17 Hz (Fig. 2B, 3C). Consequently, a linear model fails to explain the relationship between electrical activity and the MB-SWIFT fMRI signal in the awake condition (Fig. 4).

Discussion

While the exact mechanism behind MB-SWIFT fMRI contrast is not known, it has been proposed that a significant part originates from cerebral blood flow (CBF) effects¹. Therefore, our finding of a linear interrelationship between MB-SWIFT fMRI signal and electrical activity under anesthesia was expected and consistent with previous BOLD fMRI studies⁵. The discrepancy found with stimulation frequencies > 9Hz in awake rats is interesting and may point to the contribution of inhibitory activity to increased CBF through neurovascular coupling or mechanisms unexplained by hemodynamics.

Conclusion

In conclusion, MB-SWIFT fMRI provides a good proxy for neuronal activity under conditions conventionally used in experimental fMRI. This together with robustness and quietness of the approach further corroborates the use of zero echo time fMRI as an artefact free alternative to BOLD fMRI.

References

1. Lehto L, Idiyatullin D, Zhang J, et al. MB-SWIFT functional MRI during deep brain stimulation in rats. Neuroimage. 2017;159:443-48.

2. Paasonen J, Laakso H, Pirttimäki T, et al. Multi-band SWIFT enables quiet and artefact-free EEG-fMRI and awake fMRI studies in rat. Neuroimage. 2020;206:116338.

3. Laakso H, Lehto LJ, Paasonen J et al. Spinal cord fMRI with MB‐SWIFT for assessing epidural spinal cord stimulation in rats. Magnetic resonance in medicine. 2021;86(4):2137-45.

4. Paasonen J, Stenroos P, Laakso H, et al. Whole-brain studies of spontaneous behavior in head-fixed rats enabled by zero echo time MB-SWIFT fMRI. Neuroimage. 2022;250:118924.

5. Ekstrom A. How and when the fMRI BOLD signal relates to underlying neural activity: the danger in dissociation. Brain research reviews. 2010;62(2):233-44.

6. Stenroos P, Paasonen J, Salo RA, et al. Awake rat brain functional magnetic resonance imaging using standard radio frequency coils and a 3D printed restraint kit. Frontiers in neuroscience. 2018;12:548.