Kayley Marchena-Romero^1,2, Xiang Ji², Andrew Centen², Rosa Sommer^2,3, Joel Ramirez², Michael J Thrippleton⁴, Andrew Lim^2,3, Joanna M Wardlaw⁴, Sandra E Black², and Bradley J MacIntosh^1,2
1Medical Biophysics, University of Toronto, Toronto, ON, Canada, ²Sunnybrook Research Institute, Toronto, ON, Canada, ³Institute of Medical Sciences, University of Toronto, Toronto, ON, Canada, ⁴Centre for Clinical Brain Sciences, University of Edinburgh, Edinburgh, United Kingdom

Cerebrovascular reactivity (CVR) is a dynamic assessment of hemodynamic response during hypercapnia. Conventional CVR metrics provide a single estimate to summarize the dilatory response to CO₂. We propose to isolate dilation and constriction responses to a CVR challenge measured by BOLD-fMRI and assess the effect of vascular supply on vasoreactivity. We show that vascular supply affects the hemodynamic response to hypercapnia, and that this effect varies over time in different clinical cohorts. This work suggests that the extraction of individual features embedded in a CVR challenge can help characterize cerebrovascular physiology in unique clinical cohorts.

Rachael C Stickland¹, Kristina M Zvolanek^1,2, and Molly G Bright^1,2
1Physical Therapy and Human Movement Sciences, Feinberg School of Medicine, Northwestern University, Chicago, IL, United States, ²Biomedical Engineering, McCormick School of Engineering, Northwestern University, Evanston, IL, United States

We assessed the repeatability of fMRI Cerebrovascular Reactivity (CVR) estimates using CO₂ sinusoidal stimuli. We estimated BOLD-CVR amplitude and phase using a fast Fourier transform (FFT) analysis and a cross-correlation analysis (Rapidtide).  Amplitude and delay maps showed sensible regional variations for both time-points, but participant variability was notable. CVR spatial correlations between time-points were higher after averaging within cortical parcels, compared to weak voxel-wise correlations. Qualitatively, Rapidtide achieved higher between-session correlations than FFT.  GM average amplitudes agreed between time-points. At the correct spatial scale, with careful consideration of parameter scaling, this short stimulus is a viable CVR mapping option. 

Kristina M. Zvolanek^1,2, Stefano Moia³, Rachael C. Stickland¹, César Caballero-Gaudes³, and Molly G. Bright^1,2
1Physical Therapy and Human Movement Sciences, Northwestern University, Chicago, IL, United States, ²Biomedical Engineering, Northwestern University, Evanston, IL, United States, ³Basque Center on Cognition, Brain and Language, Donostia, [Gipuzkoa], Spain

Using a breath-hold task to modulate end-tidal CO2 during a BOLD fMRI scan is an established method to map hemodynamic lag. However, end-tidal CO2 measurements may not be reliable in less compliant subjects and require additional monitoring equipment. We investigated whether a related metric, respiration volume per time (RVT) can be used as a surrogate for end-tidal CO2 by directly comparing the lag values derived from each physiological measurement. Similar spatial variation was observed in healthy adult participants, albeit with some degree of proportional bias, indicating respiration belt data can be used as an alternative to map hemodynamic lag. 

Inês Esteves¹, Beatriz Raposo¹, Marta Xavier¹, Ana R. Fouto¹, Amparo Ruiz-Tagle¹, Catarina Domingos¹, Athanasios Vourvopoulos¹, Nuno A. Silva², Rita G. Nunes¹, Raquel Gil-Gouveia³, Agostinho Rosa⁴, and Patrícia Figueiredo^4
1ISR-Lisboa and Department of Bioengineering, Instituto Superior Técnico – Universidade de Lisboa, Lisboa, Portugal, ²Learning Health, Hospital da Luz, Lisboa, Portugal, ³Neurology Department, Hospital da Luz, Lisboa, Portugal, ⁴ISR-Lisboa and Department of Bioengineering, Instituto Superior Técnico – Universidade de Lisboa, Lisbon, Portugal

fMRI can be used to assess cerebrovascular reactivity by measuring BOLD changes during a breath hold task. This is assumed to be isometabolic so that BOLD changes cannot be attributed to neuronal activity. We performed simultaneous EEG-fMRI recordings during a breath-hold task to assess to what extent neuronal activity retrieved from the EEG may contribute to explain the BOLD signal variance that is usually interpreted as reactivity. We found that neuronal modulation measured by EEG band power did not explain significantly more BOLD variance across GM than PetCO2 changes, indicating that the associated fMRI measurements truly reflect cerebrovascular reactivity.

Mark Andrew Hoggarth¹, Rachael C Stickland¹, Kimberly J Hemmerling², Milap Sandhu^3,4, and Molly G Bright^1,2
1Physical Therapy and Human Movement Sciences, Northwestern University, Chicago, IL, United States, ²Biomedical Engineering, Northwestern University, Chicago, IL, United States, ³Physical Medicine and Rehabilitation, Northwestern University, Chicago, IL, United States, ⁴Shirley Ryan AbilityLab, Chicago, IL, United States

We assess the effects of an emerging intervention for rehabilitation, termed acute intermittent hypoxia (AIH), on cerebrovascular reactivity (CVR) as measured with BOLD fMRI. Lag-optimized CVR was measured pre- and post-AIH using a breath-hold task paradigm in 8 healthy participants. Five participants achieved the target drop in oxygen saturation (SpO₂) to 85%; 3 did not. Overall changes in group mean CVR were varied following AIH, slightly increasing for those who achieved the targeted SpO₂, and decreasing in those who did not. This work motivates continued study of the effects of AIH interventions on CVR.

Stefano Moia¹, Molly G Bright^2,3, and César Caballero-Gaudes^1
1Basque Center on Cognition, Brain and Language, Donostia, Spain, ²Physical Therapy and Human Movement Sciences, Feinberg School of Medicine, Northwestern University, Chicago, IL, United States, ³Biomedical Engineering, McCormick School of Engineering, Northwestern University, Evanston, IL, United States

Recent literature brought up the existence of physiological and vascular brain networks, i.e. areas showing synchronous responses to cardiac and respiratory fluctuations or hypercapnia. In this work, we show how breath-hold-induced BOLD fMRI time series from a dense mapping dataset can be parcellated into physiologically-driven networks, showing evidence of vascular contributions to the default mode network, in individuals and at the group level.

Emilie Sleight^1,2, Michael S Stringer^1,2, Madeleine Murphy³, Isla Mitchell³, Ian Marshall^1,2, Joanna M Wardlaw^1,2, and Michael J Thrippleton^1,2
1Centre for Clinical Brain Sciences, University of Edinburgh, Edinburgh, United Kingdom, ²UK Dementia Research Institute, University of Edinburgh, Edinburgh, United Kingdom, ³Edinburgh Imaging Facility Royal Infirmary of Edinburgh, University of Edinburgh, Edinburgh, United Kingdom

The BOLD response to a gas with fixed inspired CO₂ concentration is commonly used to assess the health of the cerebral vasculature. Cerebrovascular reactivity (CVR) magnitude repeatability has been assessed in previous studies, but without reporting potential physiological confounders such as heart and respiration rates. Furthermore, less is known about CVR delay repeatability. In this work, we assessed the within-day test-retest repeatability of both quantities in healthy volunteers at 3 T while measuring the heart and respiration rates. We found that both show good repeatability, though CVR delay estimates are unreliable in tissues with low contrast to noise ratio.

Colette Clare Milbourn¹ and Nicholas Paul Blockley^2
1School of Life Sciences, University of Nottingham, Nottingham, United Kingdom, ²The School of Life Sciences, University of Nottingham, Nottingham, United Kingdom

New clinical tools are needed for the diagnosis and prognosis of cerebrovascular diseases. Cerebrovascular Reactivity (CVR) is a potential marker for brain health and can be induced using stressors to the brain like high carbon dioxide (hypercapnia) challenges. These challenges can be used to map CVR. Dual-echo pseudocontinuous Arterial Spin Labelling (pCASL) and Blood Oxygen Level Dependent (BOLD) weighted imaging was combined with sinusoidally modulated and conventional block hypercapnic challenges to map CVR. Preliminary results show qualtitative improvements to pCASL based CVR maps when using sinusoidal hypercapnic challenges compared with the conventional block.