Purpose: To explore the effects of recalibration on a cardiac surgery risk of mortality (STS-ROM) model in a majority-minority population (25% white). 

Methods: 3,721 patients from 2009-20 were identified in a cardiac surgery registry at a tertiary care center in Hawai'i. Recalibration was performed by training a new model built on only this local data, with in-hospital mortality as a dependent variable, and the following three independent variable strategies: (1) only STS-ROM score (add a flat value), (2) the same independent variables as the original STS-ROM (same model, new coefficients), and (3) entirely new variable selection (local-only model). In keeping with STS methodology, logistic regression models were created with a 70:30 test:train split and optimism-adjusted LASSO variable selection when warranted. Models were evaluated via observed/expected (O/E) ratios and area under the curve (AUC). Clinical impact was evaluated in the test set by comparing classifications of the new models compared to the original STS-ROM risk classification (Low/Medium/High). During this timeframe, STS-ROM was a primary determinant of receiving surgical aortic valve replacement vs transthoracic aortic valve replacement. Thus we also calculated the percentage of aortic valve patients who would have qualified for a different procedure had we used the local model. 

Results: The STS-ROM model underestimated risk overall (O/E: 1.14). Adding a flat value (model 1) performed marginally better (O/E: 1.07) but building a new model (2) using local data and STS-ROM variables performed the best (O/E: 0.99). Model 3, which used only local data, significantly overestimated risk (O/E: 0.82). All models aside from Model 3 had similar AUC (0.82-0.83). If the best model (2) was used instead of the STS-ROM, 3.72% of all patients and 4.1% of non-white patients would have moved up in risk category. Of those already at intermediate risk, 15.6% would have been elevated to high risk. In the aortic valvular population, where STS-ROM was an especially important variable in clinical decision making, 16.7% of surgical replacements may have qualified for less-invasive procedures instead. 

Conclusion: Recalibrating national risk models using local data from our majority-minority population was superior in all models, but entirely new models without national input studies performed worse. Given that these risk models are major factors in surgical decision making, recalibration can have a significant impact on patient care. 


Surgical Risk Models in a Majority-Minority Population: The Clinical Effect of Local Recalibration

Helping is an inherently cooperative behavior, but the cognitive mechanisms underlying this behavior remain relatively underexplored. In this paper, we introduce a novel gamified paradigm for understanding a variety of cognitive behaviors associated with helping. Principals are assigned secret goals in a block-based grid (e.g., move all blue blocks to room C), and helpers can either pass their turn or make a move that could help the principal. We show that principals make useful and pragmatic first moves and helpers accumulate evidence over time before initiating a helpful move. We also introduce a preliminary set of computational models based on recursive pragmatic inference and utility maximization that attempt to account for these behavioral findings.


Help me help you: A computational model for goal inference and action planning

Decision-making models hold a vital role in the field cognitive science, serving as a means of describing and predicting human behavior. While classical models with similar assumptions are frequently favored, there is no guarantee they provide the best accounts of behavior. Here, we evaluate BEAST, a model that has demonstrated extraordinary predictive capabilities in diverse settings, but was excluded from a recent large-scale comparison of models due to its complexity. Our evaluation of the model's performance on a large collection of experiments of decisions under risk shows it provides excellent predictions in some domains. We further show how BEAST can be adapted to increase its predictive power in contextualized settings. Our results highlight the importance of considering a more inclusive approach toward complex models to deepen our understanding of the psychological mechanisms underlying human decision making behavior.


The Importance of Complex Models in Decision Making Research: An Empirical Analysis using BEAST

Face stereotypes are prevalent and consequential. In this paper, we investigate an experience sampling account of face stereotyping. In two experiments, we show that, in an experience-sampling-based Trust Game, participants in the role of an investor were much more likely to play the game with trustees with trustworthy faces than those with untrustworthy faces (Experiment 1). Crucially, this endogenous experience sampling bias has amplified the behavioral consequences of the stereotype of facial trustworthiness. In contrast, when the information was directly described with no sampling constraint for the participants, the face stereotype had a very limited effect on investment decisions (Experiment 2). We conclude by suggesting that the description-experience gap paves a promising avenue for studying sampling-based accounts in social cognition and behavior.


A description-experience gap in face stereotyping

Computational cognitive modeling is an important tool for understanding the processes that support human and animal decision-making. Choice data in sequential decision-making tasks are always noisy, and separating noise from signal can improve the quality of computational modeling. Currently, most models assume that noise is static, typically by including a parameter (e.g., uniform ε or logistic β) to estimate the noise level. However, this assumption is not guaranteed to hold -- for example, an agent can lapse into an inattentive phase for a series of trials in the middle of otherwise low-noise performance. Assuming that noise is static could bias parameter and model identification. Here, we propose a new method to dynamically infer noise in choice behavior, under a model assumption that agents can transition between two discrete latent states (for example, attentive and noisy). Using four empirical datasets with diverse behavioral and modeling features, we demonstrate that our method improves model fit and that it can be easily incorporated into existing fitting procedures, including maximum likelihood estimation and hierarchical Bayesian modeling.


A generalized method for dynamic noise inference in modeling sequential decision-making

Solving problems with others not only reduces the time required to complete a challenge but may also enable the discovery of novel strategies that qualitatively change how a problem is approached. At the dyadic level, the laboratory-based ‘shepherding task’ demonstrated that, when tasked to contain evasive agents to a centralized location, some participants discover a non-obvious but optimal strategy to solve the task. This paper quantified the interactions between participants engaged in the task using Multidimensional Cross-Recurrence Quantification Analysis (MdCRQA), applied to each participant’s gaze and hand movements. The results demonstrated that strategy discoverers exhibited greater amounts of incidental coupling than non-discoverers prior to discovery. Once discovered, the strategy reduced the strength of coupling between participants, indicating that the strategy also reduced coordination demands. Future work will investigate whether differences in problem-solving can be attributable to differences in the perceptual features participants use which scaffold the discovery of task-optimal solutions.


Incidental Coupling of Perceptual-Motor Behaviors Associated with Solution Insight during Physical Collaborative Problem-Solving

When faced with uncertainty, human observers maximize performance by integrating sensory information with learned task-relevant regularities. Does this behavior similarly occur in social settings? Here, we used a probabilistic decision-making task to explore how human behavior in an intergroup context is influenced by readily available but task-irrelevant social information (in the form of group membership) versus task-relevant reward information that could be learned over time. Across two experiments, we show that participants learned and utilized task-relevant regularities to inform their choices. We also show that human observers are not universally biased towards utilizing social information in all settings––participants learned to disregard social information when not relevant to the task at hand. However, learning about the utility of social information had a long-term influence on observers’ ability to subsequently learn and utilize available sources of information. Real-world intergroup contexts typically encompass situations and stimuli that have been previously experienced by the observer. Our findings highlight the powerful influence of learning in such contexts.


Learning and Information Use in an Intergroup Context

Recent research has identified intertemporal impulsivity as a critical cognitive variable for explaining the autocatalytic nature of socioeconomic status. But how exactly this relationship transpires has not been clearly identified, with several possible cognitive mechanisms having been proposed in the literature. We designed an experimental paradigm in which participants farmed crops under budgetary constraints and intermittently faced random resource demands. We discovered that, as a result of unpredictable resource shocks, people's preferences shifted from long-term choices to short-term ones. We also found people's self-reported sense-of-control scores to be predictive of the magnitude of their preference shifts. On the basis of these results, we argue that steep inter-temporal discounting arises as a rational adaptation to persistently experiencing long-term planning failures due to unpredictable resource shocks.


Unpredictability shortens planning horizons

Research on social motor coordination (SMC), defined as the intentional or unintentional coordination of movement between individuals in a social setting, has been linked to greater feelings of rapport and social connectedness. Here, we further investigated this relationship using a silent disco paradigm where groups of individuals listened to either the same music (the same track at the same time) or different music (different tracks at any one time). Visual information was also manipulated by having groups listened and danced with curtains separating them closed (2 minutes) or open (10 minutes). Head movements were recorded with a wireless motion tracking system, attached to the silent disco headphones. Rapport and social connectedness measures were obtained using questionnaires completed after participation in the silent disco. Results showed that groups that listened to the same music exhibited a greater degree of SMC than groups that listened to different music. Greater degrees of SMC were also observed when group members were able to see one another. Finally, greater SMC was associated with increased self-other overlap and increased perceived quality of interaction.


Social connection on the dance floor: Movement coordination in small group silent disco leads to greater self-other overlap

If we want to know whether a stack of blocks will fall, we might form our judgment by using mental simulation or simple heuristics. How do we know which strategy to use? Here, we test whether people’s strategy selection is “resource rational,” balancing their utilities and costs. We had participants judge how a balance beam will fall – a task that relies on both simulation and heuristics – and imposed visuospatial cognitive load to increase the cost of mental simulation, expecting the use of simulation to then decrease and the use of heuristics to increase. When cognitive load varied between participants, higher load led to less simulation. However, when load varied across trials for an individual participant, they instead used simulation at about the same rate as high-load-only participants. Thus strategy selection is sensitive to cognitive costs but it is also sticky, not accounting for rapidly fluctuating costs.

Strategy choice for physical reasoning is (partially) sensitive to cognitive costs

People care about others’ welfare to varying degrees, captured by the welfare tradeoff ratio (WTR), or the weight placed on another person’s welfare compared to one’s own. People can infer another person’s WTR toward themselves and reciprocate by adjusting their own WTR toward that person. However, the evolutionary origin of such a capacity is unclear. In games with perfect information, a heuristic strategy with tit-for-tat-like reciprocity is unbeatable, and the additional computation of inferring the opponent’s WTR confers no benefit. Here we show that in games with asymmetric information, where the actor has more accurate information about the payoff structure than the observer, the heuristic strategy is prone to errors from misperception, while reciprocity based on the inference of WTRs is robust. These findings suggest that asymmetric information about social decisions, a realistic modification to the game environment, may have contributed to the evolution of people’s understanding of others’ WTRs.


The evolution of reciprocity based on welfare tradeoff ratios in games with asymmetric information

Sampling assumptions --- the assumptions people make about how an example of a category or concept has been chosen --- help us learn from examples efficiently. One context where sampling assumptions are particularly important is in social contexts, where a learner needs to infer the knowledge and intentions of the information provider and vice-versa. The pedagogical sampling assumptions model describes a Bayesian account of how learners and providers should behave given different assumptions they have about the other (e.g., is the provider trying to deceive or help me? Does the learner trust me?). In this study, we tested how well this model could describe learning behaviour in the rectangle game, where a fictional information provider revealed clues about the structure of a rectangle that the learner (a participant) needed to guess. Participants received clues from either a helpful information provider, a provider who was randomly sampling clues, or one of two kinds of unhelpful providers (who could mislead but could not lie). We found that people learned efficiently and in line with model predictions when the provider was helpful and that this was the case even when no cover story was provided. However, although participants could identify that unhelpful providers were not being helpful, they struggled to learn the strategy those providers were using.


Underline digital video library

Surgical Risk Models in a Majority-Minority Population: The Clinical Effect of Local Recalibration

Help me help you: A computational model for goal inference and action planning

The Importance of Complex Models in Decision Making Research: An Empirical Analysis using BEAST

A description-experience gap in face stereotyping

A generalized method for dynamic noise inference in modeling sequential decision-making

Incidental Coupling of Perceptual-Motor Behaviors Associated with Solution Insight during Physical Collaborative Problem-Solving

Learning and Information Use in an Intergroup Context

Unpredictability shortens planning horizons

Social connection on the dance floor: Movement coordination in small group silent disco leads to greater self-other overlap

Strategy choice for physical reasoning is (partially) sensitive to cognitive costs

The evolution of reciprocity based on welfare tradeoff ratios in games with asymmetric information

Inferring the truth from deception: What can people learn from helpful and unhelpful information providers?

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES