Does digital health improve equity for Latine mental health after a disaster? Evidence from a randomized trial with 2017 and 2018 hurricane survivors.

At any given time, the environment offers a wealth of sensory stimulation and affords many potential behavioral responses, but our eyes can only look in one direction and our limbs be in one location. Narrowing down behavioral responses from potentially many to effectively one per effector (and sometimes none) requires a deliberation process that is referred to as decision. The use of sensory information to guide action choices is a core principle of all mobile animals and a fundamental drive in evolution. Estimating how quickly and consistently information can travel to support this ability is both intrinsically interesting and essential to understand the neural underpinning of all sensorimotor decisions, in healthy and diseased brains. Human brains are evolved from simpler brains capable of making basic sensory-motor decisions, in terms of selecting a behavior among different options. Indeed, one can argue that translating sensory input into behavioral responses is what brains are for (). We believe the complexities of human decisions must be evolved and extrapolated from the principles of basic sensory-motor behavior, and therefore, it is fruitful to study fast sensory-guided action as a window into understanding decision. This article focuses on the processes that take place when sensory information is used to drive behavioral responses in line with internal goals. At the core of this transformation are some neurocomputational processes, including an assessment of the sensory information with respect to internal goals, and a competition between action plans (i.e., the neuronal activity favoring alternative response options) that results in the co-determination of one response (i.e., a choice) and one reaction time (RT; if the chosen response is to move; ). These core processes are often referred to as perceptual discrimination and response selection and form part of the broader concept of decision. They start with the arrival of sensory signals and culminates with a point-of-no-return triggering one of the available motor plans. Before and after these are inevitable delays, while sensory and motor information travel to and from the neuronal populations enabling these processes. This conception superficially aligns with the way response times are modeled, that is, as the sum of decision and non-decision times (NDTs; see ; ; ; , for clear, recent articulations of this idea). But what biological processes form part of decision, and what exactly constitutes the NDT? These are important questions, as the inferences drawn about decision depend inextricably on the assumptions or inferences made about NDT. This article hopes to offer a theoretical and empirical perspective on sensorimotor delays and how these relate to the way NDT is conceptualized and estimated in the field. In a first theoretical section, we describe the principles and main assumptions of decision models and review how this literature has addressed NDT, highlighting the theoretical and empirical challenges for simplistic interpretation (; ). In a second theoretical section, we adopt a biological perspective based on recordings from monkey neurons and the oculomotor/electromyography (EMG) literature to ask when decisional processes start and end. We then put forward a broad definition of decision and NDT based on time-windows when behavioral responses (action choices and RT) are or are not influenced by competitor stimuli and explain why it is a sensible choice to fully capture and explain a range of behaviors across multiple speeded sensory-guided action tasks. This definition also allows us to measure NDT noninvasively and without model fitting. In the empirical part of the article, we then measure visuomotor deadtime (VMDT; the portion of RT when the response is not influenced by competitors) across a range of data sets, delineating its main properties. We then fit some of these data sets using the three most common decision model instantiations and contrast model NDT parameters to our empirical estimates, in order to test how well the general assumption is met that model NDTs reflect sensorimotor delays.