Modeling changes in probabilistic reinforcement learning during adolescence

In the real world, many relationships between events are uncertain and probabilistic. Uncertainty is also likely to be a more common feature of daily experience for youth because they have less experience to draw from than adults. Some studies suggest probabilistic learning may be inefficient in youths compared to adults, while others suggest it may be more efficient in youths in mid adolescence. Here we used a probabilistic reinforcement learning task to test how youth age 8-17 (N = 187) and adults age 18-30 (N = 110) learn about stable probabilistic contingencies. Performance increased with age through early-twenties, then stabilized. Using hierarchical Bayesian methods to fit computational reinforcement learning models, we show that all participants’ performance was better explained by models in which negative outcomes had minimal to no impact on learning. The performance increase over age was driven by 1) an increase in learning rate (i.e. decrease in integration time scale); 2) a decrease in noisy/exploratory choices. In mid-adolescence age 13-15, salivary testosterone and learning rate were positively related. We discuss our findings in the context of other studies and hypotheses about adolescent brain development.

Length-weight relationship of bigeye scad, Selar crumenophthalmus (Bloch, 1793): Regular vs hierarchical bayesian methods

Fisheries management or conservation requires information on length-weight relationship (LWR) for the fishing regulation and biomass estimation. This study aims to assess LWR estimation using two methods, regular and Bayesian hierarchical approached for big-eye Scad (Selar crumenophthalmus). Samples of big eye Scad were collected at several fish landings around Ambon Island from March to August 2020. Length-weight relationship measurement to obtain the parameters of W = a*Lb was tested using generalized linear model and t-test. The parameter b for monthly sampling was not significantly different (F = 0.77, df = 70, P = 0.89) and showed isometric growth b=3 (t = -1.13, df = 4, P = 0.32). Regular measurement resulted parameter log10(a) = -1.99 (±SD = 1.06) dan b = 3.06 (±SD = 0.084). Bayesian method produced parameter log10(a) = -2.07 (±SD = 0.2365) dan parameter b = 3.21 (±SD = 0.1497). Weight measurement from HB approach was significantly higher than the regular method (t = 1.65; df = 405; P <0.0001), and might produce over-estimated of weight from length data. Discrepancy of these methods was overcome by combining all information of LWR to obtain the best estimation on LWR parameters.

A Monte Carlo method to estimate cell population heterogeneity

1AbstractVariation is characteristic of all living systems. Laboratory techniques such as flow cytometry can probe individual cells, and, after decades of experimentation, it is clear that even members of genetically identical cell populations can exhibit differences. To understand whether variation is biologically meaningful, it is essential to discern its source. Mathematical models of biological systems are tools that can be used to investigate causes of cell-to-cell variation. From mathematical analysis and simulation of these models, biological hypotheses can be posed and investigated, then parameter inference can determine which of these is compatible with experimental data. Data from laboratory experiments often consist of “snapshots” representing distributions of cellular properties at different points in time, rather than individual cell trajectories. These data are not straightforward to fit using hierarchical Bayesian methods, which require the number of cell population clusters to be chosen a priori. Here, we introduce a computational sampling method named “Contour Monte Carlo” for estimating mathematical model parameters from snapshot distributions, which is straightforward to implement and does not require cells be assigned to predefined categories. Our method is appropriate for systems where observed variation is mostly due to variability in cellular processes rather than experimental measurement error, which may be the case for many systems due to continued improvements in resolution of laboratory techniques. In this paper, we apply our method to quantify cellular variation for three biological systems of interest and provide Julia code enabling others to use this method.

Reconsidering electrophysiological markers of response inhibition in light of trigger failures in the stop-signal task

We investigate the neural correlates underpinning response inhibition using a parametric ex-Gaussian model of stop-signal task performance, fit with hierarchical Bayesian methods, in a large healthy sample (N=156). The parametric model accounted for trigger failure (i.e., failures to initiate the inhibition process) and returned an SSRT estimate (SSRTEXG3) that was attenuated by ≈65ms compared to traditional non-parametric SSRT estimates (SSRTint). The amplitude and latency of the N1 and P3 event related potential components were derived for both stop-success and stop-failure trials and compared to behavioural estimates derived from traditional (SSRTint) and parametric (SSRTEXG3, trigger failure) models. Both the fronto-central N1 and P3 peaked earlier and were larger for stop-success than stop-failure trials. For stop-failure trials only, N1 peak latency correlated with both SSRT estimates as well as trigger failure and temporally coincided with SSRTEXG3, but not SSRTint. In contrast, P3 peak and onset latency were not associated with any behavioural estimates of inhibition for either trial type. While overall the N1 peaked earlier for stop-success than stop-failure trials, this effect was not found in poor task performers (i.e., high trigger failure/slow SSRT). These findings are consistent with attentional modulation of both the speed and reliability of the inhibition process, but not for poor performers. Together with the absence of any P3 onset latency effect, our findings suggest that attentional mechanisms are important in supporting speeded and reliable inhibition processes required in the stop-signal task.

Competition alters tree growth responses to climate at individual and stand scales

Understanding how climate affects tree growth is essential for assessing climate change impacts on forests but can be confounded by effects of competition, which strongly influences tree responses to climate. We characterized the joint influences of tree size, competition, and climate on diameter growth using hierarchical Bayesian methods applied to permanent sample plot data from the montane forests of Mount Rainier National Park, Washington State, USA, which are mostly comprised of Abies amabilis Douglas ex Forbes, Tsuga heterophylla (Raf.) Sarg., Pseudotsuga menziesii (Mirb.) Franco, and Thuja plicata Donn ex D. Don. Individual growth was sensitive to climate under low but not high competition, likely because tree ability to increase growth under more favorable climates (generally greater energy availability) was constrained by competition, with important variation among species. Thus, climate change will likely increase individual growth most in uncrowded stands with lower competition. However, crowded stands have more and (or) larger trees, conferring greater capacity for aggregate absolute growth increases. Due to these contrasting effects, our models predicted that climate change will lead to greater stand-scale growth increases in stands with medium compared with low crowding but similar increases in stands with medium and high crowding. Thus, competition will mediate the impacts of climate change on individual- and stand-scale growth in important but complex ways.