RESUMEN
Cerebral blood flow, cerebral stiffness (CS) and intracranial pressure are tightly linked variables of cerebrovascular reactivity and cerebral autoregulation. Transtemporal ultrasound time-harmonic elastography was used for rapid measurement of CS changes in 10 volunteers before, during and after administration of a gas mixture of 95% O2 and 5% CO2 (carbogen). Within the first 2.2 ± 2.0 min of carbogen breathing, shear wave speed determined as a surrogate parameter of CS increased from 1.57 ± 0.04 to 1.66 ± 0.05 m/s (p < 0.01) in synchrony with end-tidal CO2 while post-hypercapnic CS recovery was delayed by 2.7 ± 1.4 min in relation to end-tidal CO2. Our results indicate that CS is highly sensitive to changes in CO2 levels of inhaled air. Possible mechanisms underlying the observed CS changes might be associated with cerebrovascular reactivity, cerebral blood flow adaptation and intracranial regulation, all of which are potentially relevant for future diagnostic applications of transtemporal time-harmonic elastography in a wide spectrum of neurologic diseases.
Asunto(s)
Hipercapnia/patología , Rigidez Vascular/efectos de los fármacos , Adulto , Dióxido de Carbono/efectos adversos , Circulación Cerebrovascular/efectos de los fármacos , Diagnóstico por Imagen de Elasticidad/métodos , Femenino , Humanos , Hipercapnia/diagnóstico por imagen , Masculino , Adulto JovenRESUMEN
Current learning machines have successfully solved hard application problems, reaching high accuracy and displaying seemingly intelligent behavior. Here we apply recent techniques for explaining decisions of state-of-the-art learning machines and analyze various tasks from computer vision and arcade games. This showcases a spectrum of problem-solving behaviors ranging from naive and short-sighted, to well-informed and strategic. We observe that standard performance evaluation metrics can be oblivious to distinguishing these diverse problem solving behaviors. Furthermore, we propose our semi-automated Spectral Relevance Analysis that provides a practically effective way of characterizing and validating the behavior of nonlinear learning machines. This helps to assess whether a learned model indeed delivers reliably for the problem that it was conceived for. Furthermore, our work intends to add a voice of caution to the ongoing excitement about machine intelligence and pledges to evaluate and judge some of these recent successes in a more nuanced manner.