The aim of this study is twofold: on the one hand, to determine how visual space, as assessed by exocentric distance estimates, is related to physical space. On the other hand, to determine the structure of visual space as assessed by exocentric distance estimates. Visual space was measured in three environments: (a) points located in a 2-D frontoparallel plane, covering a range of distances of 20 cm; (b) stakes placed in a 3-D virtual space (range ≈ 330 mm); and (c) stakes in a 3-D outdoors open field (range = 45 m). Observers made matching judgments of distances between all possible pairs of stimuli, obtained from 16 stimuli (in a regular squared 4 × 4 matrix). Two parameters from Stevens' power law informed us about the distortion of visual space: its exponent and its coefficient of determination (R2). The results showed a ranking of the magnitude of the distortions found in each experimental environment, and also provided information about the efficacy of available visual cues of spatial layout. Furthermore, our data are in agreement with previous findings showing systematic perceptual errors, such as the further the stimuli, the larger the distortion of the area subtended by perceived distances between stimuli. Additionally, we measured the magnitude of distortion of visual space relative to physical space by a parameter of multidimensional scaling analyses, the RMSE. From these results, the magnitude of such distortions can be ranked, and the utility or efficacy of the available visual cues informing about the space layout can also be inferred.

In Natural Language Processing (NLP) and related fields, computational models of mental health statuses prediction aim to detect early signs of mental health issues based on an individual’s behaviour on social media. Models of this kind, which are mostly devoted to depression disorders and to the English language, present many open research questions. First, since context-free social media data are prone to noise, the task may involve processing large amounts of data with little or no relation to mental health, which may hinder both model efficiency and accuracy. Second, existing work has been largely devoted to text processing, even though social media also include a wide range of non-textual information, which may be predictors of mental health in their own right. Finally, existing models are usually validated in a single domain, often involving one dataset of a particular text genre and language, and it is not clear whether their results may generalise to other scenarios. Based on these observations, the present work introduces a number of computational models for the prediction of depression and anxiety disorder using large language models (LLM) to handle noisy, context-free social media data. These combine textual and non-textual information with the aid of mixture of experts (MoE), and are evaluated in both the Twitter/X domain in Portuguese and in the Reddit domain in English using both machine learning metrics and human assessment provided by mental health specialists. Results show a number of improvements over the previous work and suggest new lines of investigation.

The main purpose of the present study was to investigate whether in natural environment, using very large physical distances, there is a trend to overconstancy for distance estimates during development. One hundred and twenty-nine children aged 5 to 13 years old and twenty-one adults (in a control group), participated as observers. The observer's task was to bisect egocentric distances, ranging from 1.0 to 296.0 m, presented in a large open field. The analyses focused on two parameters, constant errors and variable errors, such as measuring accuracy and precision, respectively. A third analysis focused on the developmental pattern of shifts in constancy as a function of age and range of distances. Constant error analysis showed that there are two relevant parameters for accuracy, age, and range of distances. For short distances, there are three developmental stages: 5-7 years, when children have unstable responses, 7-11, underconstancy, and 13 to adulthood, when accuracy is reached. For large distances, there is a two-stage development: 5-11 years, with severe underconstancy, and beyond this age, with mild underconstancy. Variable errors analyses indicate that precision is noted for 7 year-old children, independently of the range of distances. The constancy analyses indicated that there is a shift from constancy (or slightly overconstancy) to underconstancy as a function of physical distance for all age groups. The age difference is noted in the magnitude of underconstancy that occurs in larger distances, where adults presented lower levels of underconstancy than children. The present data were interpreted as due to a developmental change in cognitive processing rather than to changes in visual space perception.