Digital Theses Archive


Tesi etd-03252022-090725

Type of thesis
functional MRI of the brainstem in Cheyne-Stokes respiration: a methodological and translational investigation
Scientific disciplinary sector
Istituto di Scienze della Vita - TRANSLATIONAL MEDICINE 2018
  • componenti indipendenti
  • respiro
  • risonanza magnetica
  • chemocezione
  • ipercapnia
  • hypercapnia
  • breath hold
  • independent component analysis
  • cheyne-stokes respiration
  • breathing
  • respiration
  • chemoreception
  • fMRI
  • apnea
  • respiro periodico
Exam session start date
The central control of breathing is a complex system in which several units concur to regulate the respiration pattern and maintain the homeostasis in arterial blood PO2, PCO2 and pH. The main actors of this intricated system are located in the brainstem and are subject to malfunctions in the presence of pathologies involving the autonomic system. Heart Failure (HF) is a complex syndrome that introduces a general unbalance in the autonomic system, possibly affecting the central nervous system. One of the main determinants in the prognosis and progression of HF is Cheyne Stokes Respiration (CSR), a central abnormal breathing pattern characterized by the alternation between states of hyperventilation and states of apnea. Targeting CSR is of huge importance for the scientific community to reduce the hospitalization and mortality rates of HF. Unfortunately, the pathophysiology of CSR is still understudied, and the deep brain structures deputed to breathing control and CO2 chemoreception are hard to noninvasively access. Functional Magnetic Resonance Imaging (fMRI) is the best candidate for the investigation of neural alterations at the base of CSR, as it offers good spatio-temporal resolution even in the deepest brain regions. <br><br>In this work, fMRI is used to study the central system of breathing control, and in particular the role of central chemoreceptors in the physiopathology of CSR, employing an analytic approach based on the work of Beissner et al. on masked independent component analysis. We identify four steps needed to reach this goal. First, to define a methodology to face the challenges posed by the study of the brainstem with fMRI. This means to evaluate the adaptability of existing approaches to the study of the central control of breathing, in particular for what regards physiological noise correction and signal modelling. Second, to define a study design and an analysis approach for the characterization of the central control of breathing in healthy subjects. Third, to acquire and analyze images from HF and CSR patients in order to investigate the alterations at the base of CSR. Fourth and last, to target the lack of tools for the analysis of brainstem structures, focusing on the development of a novel brainstem atlas. Following this plan, we expose relevant considerations on the correct approach to physiological noise correction in breath hold studies on neural activity related to breathing control. <br><br>In the phase of methodology development, a novel approach to the analysis of specific and non-specific BOLD-signal changes related to end-tidal CO2 in breath-hold fMRI studies is proposed. The approach employs multiple-order nonlinear predictors for end-tidal CO2 to model region-dependent nonlinear input-output relationships hypothesized in literature and possibly playing a crucial role in disentangling noise. The goal is to explore the effects of the standard retrospective image-based correction procedure for physiological noise on the estimated BOLD response. Through this framework, we also highlight cortical and subcortical activations elicited on healthy subjects by breath-hold induced hypercapnia and we describe them in terms of response complexity and delay. <br><br>We then describe the activations extracted with independent components analysis in healthy subjects and patients, discussing possible alterations in the deepest brain structures deputed to chemosensation. In addition, we shed new light on the recruitment of arousal and alarming systems in presence of hypercapnic events and on the mechanisms at the base of the exchange of information with higher cortical areas. Activations are localized with high accuracy thanks to our contribution to the development, validation and testing of a novel 7 Tesla probabilistic MRI atlas of the brainstem (Brainstem Navigator Atlas). On patients expressing a significant occurrence of diurnal apneas (AHId≥25), we hypothesize an increase in the chemosensitive response of serotonergic nuclei in the medullary raphe. We also display chemosensitive activity in a caudo-ventral region of the medulla, possibly related to understudied chemoreceptive regions observed in animal studies. We provide an indication for a concurrent weakening of the peripheral chemosensitive input and for a CO2-dependent activation of hypothalamus specific to the arousal context of CSR patients. We confirm the missing recruitment of chemosensitive alarming systems for hypercapnia in the dorsal pons, possibly at the base of the unconscious periodicity of the breathing pattern in CSR patients. Finally, we describe the spatial and temporal characteristics of a cortico-brainstem pathway linking cortical and subcortical autonomic networks. <br><br>Although we recognize the need to access these deep brainstem regions with more powerful technology, we believe our conclusions represent a first promising, detailed observation of the alterations determining the derangement of HF into CSR.