Fast diffusion fMRI signals in the Superior Colliculus during the negative BOLD regime
Poster No:
2434
Submission Type:
Abstract Submission
Authors:
Rita Gil1, Noam Shemesh1
Institutions:
1Champalimaud Research, Champalimaud Foundation, Lisbon, Lisbon
First Author:
Co-Author:
Introduction:
Positive BOLD Responses (PBRS) are known to correlate with increases in local field potentials (LFPs) and multi-unit activity (MUA) signals1,2. However, there is still debate on the biological underpinnings of Negative BOLD Responses (NBRs), with evidence pointing to neuronal suppression as the most probable biological scenario3,4.
Recently, we employed a visual paradigm capable of modulating visual pathway BOLD responses from activation to suppression5 (the latter achieved at high stimulation frequencies) and showed that they corresponded to PBR->NBR transitions. To better understand and characterize the activation / suppression regimes, we harnessed diffusion functional MRI (dfMRI), whose fast responses were previously proposed to rely on neuromechanical coupling6, thereby making them more specific to neural activity and less prone to blood vessel contamination7. Our findings reinforce the potential of dfMRI as a better tracker of neural activity.
Recently, we employed a visual paradigm capable of modulating visual pathway BOLD responses from activation to suppression5 (the latter achieved at high stimulation frequencies) and showed that they corresponded to PBR->NBR transitions. To better understand and characterize the activation / suppression regimes, we harnessed diffusion functional MRI (dfMRI), whose fast responses were previously proposed to rely on neuromechanical coupling6, thereby making them more specific to neural activity and less prone to blood vessel contamination7. Our findings reinforce the potential of dfMRI as a better tracker of neural activity.
Methods:
Study in Long-Evans rats under medetomidine sedation in accordance with European Directive 2010/63.
Stimulation: A blue LED (λ=470nm) was used for binocular stimulation at 1 and 25 Hz (Fig.1A). The paradigm is shown in Fig.1B.
MRI: Data was acquired in a 9.4T BioSpec scanner at 95%O2 with an isotropically diffusion weighted SE-EPI sequence (TE/TR=37.3/750ms, FOV=18x18mm², resolution=250x250μm², slice thickness=1.5mm, b-value=0 and 1.5 ms/μm², two consecutive waveforms totalling 24.73 ms, tacq=6min and 45s) – schematic shown in Fig.1C. One tilted slice capturing the entire visual pathway was used (Fig.1D).
MRI Pre-processing: Outlier, MP-PCA denoising with a 5x5 kernel, motion and slice-timing correction and isotropic smoothing. Runs were detrended with a polynomial fit to rest periods. Apparent diffusion coefficients (ADC) were calculated per timepoint: ADC=-(1/b)*ln(Sb⁄S0).
Stimulation: A blue LED (λ=470nm) was used for binocular stimulation at 1 and 25 Hz (Fig.1A). The paradigm is shown in Fig.1B.
MRI: Data was acquired in a 9.4T BioSpec scanner at 95%O2 with an isotropically diffusion weighted SE-EPI sequence (TE/TR=37.3/750ms, FOV=18x18mm², resolution=250x250μm², slice thickness=1.5mm, b-value=0 and 1.5 ms/μm², two consecutive waveforms totalling 24.73 ms, tacq=6min and 45s) – schematic shown in Fig.1C. One tilted slice capturing the entire visual pathway was used (Fig.1D).
MRI Pre-processing: Outlier, MP-PCA denoising with a 5x5 kernel, motion and slice-timing correction and isotropic smoothing. Runs were detrended with a polynomial fit to rest periods. Apparent diffusion coefficients (ADC) were calculated per timepoint: ADC=-(1/b)*ln(Sb⁄S0).
Results:
As observed for the non-diffusion-weighted experiments, dfMRI responses in the superior colliculus (SC) and visual cortex (VC) show strong modulations with increasing stimulation frequency (Fig.2A). It is interesting to observe that dfMRI signals appear more localised to the superficial SC layers which correspond to the visual layers.
Moreover, SC dfMRI time profiles show increased CNR for both BOLD regimes (Fig.2B) along with the sharp onset and offset peaks observed for the non-diffusion-weighted time profiles.
Previous studies have reported ADC decreases in active brain regions; however the source of such ADC changes is not completely known with transient cell swelling/beading, increased turtuosity or a transition of water molecules from a faster to slower pool presenting as good candidates. While similar results are observed in this study for the PBR regime (Fig.2C), an ADC increase is observed for the NBR regime. In Fig.2D we propose a mechanism behind the observed ADC changes: A basal transient cell swelling (including dendritic and axonal beading as well as increased button size) exists during rest. These events dramatically increase upon neuronal activation leading to the already described ADC decrease due to increased extracellular tortuosity. Our previous work showing neuronal suppression during SC NBRs points to the hypothesis of reduced basal cell swelling upon neuronal suppression leading to increased ADC percentage resulting from decreased extracellular tortuosity.
Moreover, SC dfMRI time profiles show increased CNR for both BOLD regimes (Fig.2B) along with the sharp onset and offset peaks observed for the non-diffusion-weighted time profiles.
Previous studies have reported ADC decreases in active brain regions; however the source of such ADC changes is not completely known with transient cell swelling/beading, increased turtuosity or a transition of water molecules from a faster to slower pool presenting as good candidates. While similar results are observed in this study for the PBR regime (Fig.2C), an ADC increase is observed for the NBR regime. In Fig.2D we propose a mechanism behind the observed ADC changes: A basal transient cell swelling (including dendritic and axonal beading as well as increased button size) exists during rest. These events dramatically increase upon neuronal activation leading to the already described ADC decrease due to increased extracellular tortuosity. Our previous work showing neuronal suppression during SC NBRs points to the hypothesis of reduced basal cell swelling upon neuronal suppression leading to increased ADC percentage resulting from decreased extracellular tortuosity.
Conclusions:
Our results not only provide insight into the ongoing debate on the nature of negative BOLD signals but also open doors to investigate the biological mechanisms behind the measured ADC changes upon neuronal activation/suppression while highlighting the potential of dfMRI as a complementary and more cellular-oriented functional contrast mechanism.
Modeling and Analysis Methods:
Activation (eg. BOLD task-fMRI) 2
Neuroanatomy, Physiology, Metabolism and Neurotransmission:
Subcortical Structures
Novel Imaging Acquisition Methods:
BOLD fMRI
Diffusion MRI
Non-BOLD fMRI 1
Keywords:
fMRI CONTRAST MECHANISMS
FUNCTIONAL MRI
MRI
Sub-Cortical
Vision
Other - non-BOLD contrasts
1|2Indicates the priority used for review
Provide references using author date format
2. Logothetis, N. K., et al. (2001), 'Neurophysiological investigation of the basis of the fMRI signal.' Nature 412, 150–157.
3. Shmuel, A., et al. (2006), 'Negative functional MRI response correlates with decreases in neuronal activity in monkey visual area V1.' Nature Neuroscience 9, 569–577.
4. Devor, A. et al. (2007), 'Suppressed Neuronal Activity and Concurrent Arteriolar Vasoconstriction May Explain Negative Blood Oxygenation Level-Dependent Signal.', Journal of Neuroscience 27, 4452–4459.
5. Gil, Rita, et al. (2022), 'Activation/suppression balances in rat Superior Colliculus encode the visual continuity illusion.' bioRxiv: 2022-11.
6. Nunes, D., et al. (2021) 'A rapid-onset diffusion functional MRI signal reflects neuromorphological coupling dynamics.', Neuroimage 231, 117862.
7. Kim, S.-G.,et al. (1999), 'Diffusion-weighted spin-echo fMRI at 9.4 T: Microvascular/tissue contribution to BOLD signal changes.', Magnetic Resonance in Medicine 42, 919–928.