3151
Locus coeruleus is associated with brain state switching1Department of Bioengineering, University of California, Riverside, Riverside, CA, United States, 2Department of Psychology, University of California, Riverside, Riverside, CA, United States, 3Center for Advanced Neuroimaging, University of California, Riverside, Riverside, CA, United States, 4Department of Cognitive Sciences, University of California, Irvine, Irvine, CA, United States
Synopsis
Locus coeruleus (LC) is involved in attention and other brain functions. We ascertained how LC activity up-regulation affects brain states derived from a hidden Markov model. Upon up-regulating LC activity, we found increased pupil size during state switching. This may indicate an increase in LC activation during the state switching process. Specifically, significant interactions between state and condition were observed when transitioning into the default mode network, but not into attention networks. The lack of interaction may be due to LC already being active during transitions into the attention network.
Introduction
The locus coeruleus (LC) is the primary source of norepinephrine for the brain and is deeply involved in attention and task performance1,2. When LC activity is up-regulated, norepinephrine is released causing pupils to dilate thereby providing a means to indirectly measure LC activity3,4. Despite observed correlations between LC engagement and attention, its impact on underlying mechanisms driving changes in network dynamics as a function of arousal are not well understood5.Squeezing an exercise-ball has been shown to increase arousal via the LC-noradrenergic system6,7. We modified this task to create a pseudo-resting state paradigm for ascertaining how LC activity up-regulation affects brain states. In this abstract, we explore the relationship between LC activation and brain states derived from a hidden Markov model (HMM).
Methods
Twenty-six participants (17 females, mean age 25 years ± 4 years) enrolled in this study. MRI data were called on a Siemens 3T Prisma MRI scanner with a 64 channel receive-only head coil. fMRI data were collected using a 2D-EPI sequence (echo time (TE)/repetition time (TR)=32/2000 ms, flip angle=77°, and voxel size=2×2×3mm3,slices=52) while pupillometry data were collected concurrently with a TrackPixx system (VPixx, Montreal, Canada). Anatomic images from a MP-RAGE sequence (TE/TR/inversion time=3.02/2600/800 ms, flip angle=8°, voxel size=0.8×0.8×0.8 mm3) were used for registration from subject space to common space.The pseudo-resting state paradigm used in this analysis consisted of five resting state blocks separated by five squeeze blocks as illustrated in Fig. 17. Each subject underwent two scanning sessions on different days: one where they squeezed an exercise-ball (active session), and one where they brought their arm up to their chest but refrained from squeezing (sham session). Pupillometry data were preprocessed via the ET-remove-artifacts toolbox (https://github.com/RingoHHuang/ET-remove-artifacts), normalized by dividing by the mean of that session’s first resting state block, and downsampled to match the fMRI’s temporal resolution.
A hidden Markov model (HMM) was applied to z-scored blood oxygenation level dependent-signal acquired in regions of interest from the default mode network (DMN), fronto-parietal control network (FPCN), dorsal attention network (DAN), salience network (SN), and LC, which were concatenated across all subjects and conditions8–10. The HMM returned mean activation states, a latent state sequence for every subject and condition, a transition probability matrix, and a covariance matrix for every state. To ensure that these activation states are robust and reproducible, the HMM was run with five sets of random initial values and compared to the output from a model with uniform initial probability distribution. This was repeated for models containing three to nine hidden states and the optimal number of states describing the data was determined.
After identifying a switch in subjects’ state sequences, the difference between the pupil size two TRs before the switch and the first TR after the switch was calculated. This calculation was contingent on the subject remaining in that same state for two TRs before or after the identified switch to ensure that the subject settled into a stable state. 2 (active vs. sham) x 3 (transitioning to a state) repeated measure ANOVAs were then performed to assess significance of subjects transitioning into a specific state.
Results
HMMs appropriated with five or more states showed repeated activation patterns and were discarded; thus, we determined that a 4-state model best fit our dataset. The latent brain states describing our dataset are seen in Fig. 2 where States 1 (S1) and 2 (S2) correspond to DMN and DAN/SN, respectively. State 3 (S3) corresponds to a transient state acting as a hub when subjects switch between other states11,12. State 4 (S4) is prevalent during the squeeze blocks of the paradigm.Pupil dilation during state switching showed that, on average, pupil size increases during the active session compared to the sham session (p=0.0099) (Fig. 3). The omnibus ANOVAs found significant interactions between state and condition when transitioning to S1 (p=0.042) but not S2 (p=0.228) or S4 (p=0.351). No subject transitioned from S4 into S3, so a corresponding ANOVA could not be performed. Based on LC’s relationship with attention, we hypothesized that pupil dilation occurs when transitioning to the attention-dominant state. Post-hoc one-tailed paired t-tests found that pupil size is greater during active compared to sham sessions when subjects transition into S2 from S1 (p=0.0395), S3 (p=0.0162), or S4 (p=0.0485). Conversely, a contraction was observed when transitioning into S1 from S2 (p=0.0042) or S4 (p=0.0253) in the active session (Fig. 4).
Discussion
Norepinephrine is released during LC activation which causes pupil dilation3,4. Upon up-regulating LC activity, we found increased pupil size during state switching, which indicates increased LC activation during the state switching process. ANOVAs found significant interactions when transitioning into S1, but not into S2. This lack of interaction in S2 may suggest that LC is already up-regulated when transitioning into the attention-network dominant state regardless of session, but may only be up-regulated during transitions into S1 in the active session. However, further work is needed to verify this hypothesis. Overall, LC may be associated with state switching, potentially providing insight into its relationship with latent brain states.Acknowledgements
We humbly thank Ms. Chelsea Evelyn for collecting the fMRI data presented here, and Dr. Xu “Jerry” Chen for his help in collecting the pupillometry data and in implementing the task paradigm.References
1. Aston-Jones, G. & Cohen, J. D. An Integrative Theory of Locus Coeruleus-Norepinephrine Function: Adaptive Gain and Optimal Performance. Annual Review of Neuroscience vol. 28 403–450 (2005).
2. Song, A. H. et al. Pharmacological Modulation of Noradrenergic Arousal Circuitry Disrupts Functional Connectivity of the Locus Ceruleus in Humans. J. Neurosci. 37, 6938–6945 (2017).
3. Murphy, P. R., O’Connell, R. G., O’Sullivan, M., Robertson, I. H. & Balsters, J. H. Pupil diameter covaries with BOLD activity in human locus coeruleus. Hum. Brain Mapp. 35, 4140–4154 (2014).
4. Joshi, S., Li, Y., Kalwani, R. M. & Gold, J. I. Relationships between Pupil Diameter and Neuronal Activity in the Locus Coeruleus, Colliculi, and Cingulate Cortex. Neuron 89, 221–234 (2016).
5. Sara, S. J. The locus coeruleus and noradrenergic modulation of cognition. Nat. Rev. Neurosci. 10, 211–223 (2009).
6. Nielsen, S. E. & Mather, M. Comparison of two isometric handgrip protocols on sympathetic arousal in women. Physiol. Behav. 142, 5–13 (2015).
7. Mather, M. et al. Isometric exercise facilitates attention to salient events in women via the noradrenergic system. Neuroimage 210, 116560 (2020).
8. Raichle, M. E. The Restless Brain. Brain Connectivity vol. 1 3–12 (2011).
9. Deshpande, G., Santhanam, P. & Hu, X. Instantaneous and causal connectivity in resting state brain networks derived from functional MRI data. Neuroimage 54, 1043–1052 (2011).
10. Langley, J., Huddleston, D. E., Liu, C. J. & Hu, X. Reproducibility of locus coeruleus and substantia nigra imaging with neuromelanin sensitive MRI. MAGMA 30, 121–125 (2017).
11. Chen, S., Langley, J., Chen, X. & Hu, X. Spatiotemporal Modeling of Brain Dynamics Using Resting-State Functional Magnetic Resonance Imaging with Gaussian Hidden Markov Model. Brain Connect. 6, 326–334 (2016).
12. Stevner, A. B. A. et al. Discovery of key whole-brain transitions and dynamics during human wakefulness and non-REM sleep. Nat. Commun. 10, 1035 (2019).
Figures
