Human Representation Learning.

The central theme of this review is the dynamic interaction between information selection and learning. We pose a fundamental question about this interaction: How do we learn what features of our experiences are worth learning about? In humans, this process depends on attention and memory, two cognitive functions that together constrain representations of the world to features that are relevant for goal attainment. Recent evidence suggests that the representations shaped by attention and memory are themselves inferred from experience with each task. We review this evidence and place it in the context of work that has explicitly characterized representation learning as statistical inference. We discuss how inference can be scaled to real-world decisions by approximating beliefs based on a small number of experiences. Finally, we highlight some implications of this inference process for human decision-making in social environments.

Affect-congruent attention modulates generalized reward expectations.

Positive and negative affective states are respectively associated with optimistic and pessimistic expectations regarding future reward. One mechanism that might underlie these affect-related expectation biases is attention to positive- versus negative-valence features (e.g., attending to the positive reviews of a restaurant versus its expensive price). Here we tested the effects of experimentally induced positive and negative affect on feature-based attention in 120 participants completing a compound-generalization task with eye-tracking. We found that participants' reward expectations for novel compound stimuli were modulated in an affect-congruent way: positive affect induction increased reward expectations for compounds, whereas negative affect induction decreased reward expectations. Computational modelling and eye-tracking analyses each revealed that these effects were driven by affect-congruent changes in participants' allocation of attention to high- versus low-value features of compounds. These results provide mechanistic insight into a process by which affect produces biases in generalized reward expectations.

Best Practice Alert to Promote Screening for Primary Aldosteronism Among People With Apparent Treatment-Resistant Hypertension.

OBJECTIVE: Guidelines recommend screening all individuals with resistant hypertension for primary aldosteronism (PA) but less than 2% are screened. We aimed to develop a noninterruptive Best Practice Alert (BPA) to assess if its implementation in the electronic health record improved PA screening rates among individuals with apparent treatment-resistant hypertension (aTRH). METHODS: We implemented a noninterruptive BPA on 9/17/2022 at our ambulatory primary care, endocrinology, nephrology, and cardiology clinics. We assessed clinical parameters of people with aTRH before (9/17/2021-9/16/2022) and after (9/17/2022-9/16/2023) the BPA was implemented. The noninterruptive BPA embedded with an order set identified people with aTRH and recommended screening for PA if it was not previously performed. RESULTS: There were 10 944 and 11 463 people with aTRH who attended office visits during the 12 months before and after the BPA implementation, respectively. There were no statistically significant differences in median age (P = .096), sex (P = .577), race (P = .753), and ethnicity (P = .472) between the pre- and post-BPA implementation groups. There was a significant increase in PA screening orders placed (227 [2.1%] vs 476 [4.2%], P < .001) and PA screening labs performed (169 [1.5%] vs 382 [3.3, P < .001) after BPA implementation. PA screening tests were positive in 26% (44/169) and 23% (88/382) of people in the pre- and post-BPA groups, respectively (P = .447). CONCLUSION: Implementation of a real-time electronic health record BPA doubled the screening rate for PA among people with aTRH; however, the overall screening rate was low.

Intact Reinforcement Learning But Impaired Attentional Control During Multidimensional Probabilistic Learning in Older Adults.

To efficiently learn optimal behavior in complex environments, humans rely on an interplay of learning and attention. Healthy aging has been shown to independently affect both of these functions. Here, we investigate how reinforcement learning and selective attention interact during learning from trial and error across age groups. We acquired behavioral and fMRI data from older and younger adults (male and female) performing two probabilistic learning tasks with varying attention demands. Although learning in the unidimensional task did not differ across age groups, older adults performed worse than younger adults in the multidimensional task, which required high levels of selective attention. Computational modeling showed that choices of older adults are better predicted by reinforcement learning than Bayesian inference, and that older adults rely more on reinforcement learning-based predictions than younger adults. Conversely, a higher proportion of younger adults' choices was predicted by a computationally demanding Bayesian approach. In line with the behavioral findings, we observed no group differences in reinforcement-learning related fMRI activation. Specifically, prediction-error activation in the nucleus accumbens was similar across age groups, and numerically higher in older adults. However, activation in the default mode was less suppressed in older adults in for higher attentional task demands, and the level of suppression correlated with behavioral performance. Our results indicate that healthy aging does not significantly impair simple reinforcement learning. However, in complex environments, older adults rely more heavily on suboptimal reinforcement-learning strategies supported by the ventral striatum, whereas younger adults use attention processes supported by cortical networks.SIGNIFICANCE STATEMENT Changes in the way that healthy human aging affects how we learn to optimally behave are not well understood; it has been suggested that age-related declines in dopaminergic function may impair older adult's ability to learn from reinforcement. In the present fMRI experiment, we show that learning and nucleus accumbens activation in a simple unidimensional reinforcement-learning task was not significantly affected by age. However, in a more complex multidimensional task, older adults showed worse performance and relied more on reinforcement-learning strategies than younger adults, while failing to disengage their default-mode network during learning. These results imply that older adults are only impaired in reinforcement learning if they additionally need to learn which dimensions of the environment are currently important.

Association of Biotin Ingestion With Performance of Hormone and Nonhormone Assays in Healthy Adults.

Importance: Biotinylated antibodies and analogues, with their strong binding to streptavidin, are used in many clinical laboratory tests. Excess biotin in blood due to supplemental biotin ingestion may affect biotin-streptavidin binding, leading to potential clinical misinterpretation. However, the degree of interference remains undefined in healthy adults. Objective: To assess performance of specific biotinylated immunoassays after 7 days of ingesting 10 mg/d of biotin, a dose common in over-the-counter supplements for healthy adults. Design, Setting, and Participants: Nonrandomized crossover trial involving 6 healthy adults who were treated at an academic medical center research laboratory. Exposure: Administration of 10 mg/d of biotin supplementation for 7 days. Main Outcomes and Measures: Analyte concentrations were compared with baseline (day 0) measures on the seventh day of biotin treatment and 7 days after treatment had stopped (day 14). The 11 analytes included 9 hormones (ie, thyroid-stimulating hormone, total thyroxine, total triiodothyronine, free thyroxine, free triiodothyronine, parathyroid hormone, prolactin, N-terminal pro-brain natriuretic peptide, 25-hydroxyvitamin D) and 2 nonhormones (prostate-specific antigen and ferritin). A total of 37 immunoassays for the 11 analytes were evaluated on 4 diagnostic systems, including 23 assays that incorporated biotin and streptavidin components and 14 assays that did not include biotin and streptavidin components and served as negative controls. Results: Among the 2 women and 4 men (mean age, 38 years [range, 31-45 years]) who took 10 mg/d of biotin for 7 days, biotin ingestion-associated interference was found in 9 of the 23 (39%) biotinylated assays compared with none of the 14 nonbiotinylated assays (P = .007). Results from 5 of 8 biotinylated (63%) competitive immunoassays tested falsely high and results from 4 out of 15 (27%) biotinylated sandwich immunoassays tested falsely low. Conclusions and Relevance: In this preliminary study of 6 healthy adult participants and 11 hormone and nonhormone analytes measured by 37 immunoassays, ingesting 10 mg/d of biotin for 1 week was associated with potentially clinically important assay interference in some but not all biotinylated assays studied. These findings should be considered for patients taking biotin supplements before ordering blood tests or when interpreting results. Trial Registration: clinicaltrials.gov Identifier: NCT03034707.

Reinforcement learning in multidimensional environments relies on attention mechanisms.

In recent years, ideas from the computational field of reinforcement learning have revolutionized the study of learning in the brain, famously providing new, precise theories of how dopamine affects learning in the basal ganglia. However, reinforcement learning algorithms are notorious for not scaling well to multidimensional environments, as is required for real-world learning. We hypothesized that the brain naturally reduces the dimensionality of real-world problems to only those dimensions that are relevant to predicting reward, and conducted an experiment to assess by what algorithms and with what neural mechanisms this "representation learning" process is realized in humans. Our results suggest that a bilateral attentional control network comprising the intraparietal sulcus, precuneus, and dorsolateral prefrontal cortex is involved in selecting what dimensions are relevant to the task at hand, effectively updating the task representation through trial and error. In this way, cortical attention mechanisms interact with learning in the basal ganglia to solve the "curse of dimensionality" in reinforcement learning.

Statistical computations underlying the dynamics of memory updating.

Psychophysical and neurophysiological studies have suggested that memory is not simply a carbon copy of our experience: Memories are modified or new memories are formed depending on the dynamic structure of our experience, and specifically, on how gradually or abruptly the world changes. We present a statistical theory of memory formation in a dynamic environment, based on a nonparametric generalization of the switching Kalman filter. We show that this theory can qualitatively account for several psychophysical and neural phenomena, and present results of a new visual memory experiment aimed at testing the theory directly. Our experimental findings suggest that humans can use temporal discontinuities in the structure of the environment to determine when to form new memory traces. The statistical perspective we offer provides a coherent account of the conditions under which new experience is integrated into an old memory versus forming a new memory, and shows that memory formation depends on inferences about the underlying structure of our experience.

Naturalistic reinforcement learning.

Humans possess a remarkable ability to make decisions within real-world environments that are expansive, complex, and multidimensional. Human cognitive computational neuroscience has sought to exploit reinforcement learning (RL) as a framework within which to explain human decision-making, often focusing on constrained, artificial experimental tasks. In this article, we review recent efforts that use naturalistic approaches to determine how humans make decisions in complex environments that better approximate the real world, providing a clearer picture of how humans navigate the challenges posed by real-world decisions. These studies purposely embed elements of naturalistic complexity within experimental paradigms, rather than focusing on simplification, generating insights into the processes that likely underpin humans' ability to navigate complex, multidimensional real-world environments so successfully.

The Impact of Multicultural Interfacility Video Case Conference: A Novel Education Model After the COVID Pandemic.

Context: The COVID-19 pandemic challenged undertaking gradual educational activities for residency and fellowship trainees. However, recent technological advances have enabled broadening active learning opportunities through international online conferences. Objective: The format of our international online endocrine case conference, launched during the pandemic, is introduced. The objective impact of this program on trainees is described. Methods: Four academic facilities developed a semiannual international collaborative endocrinology case conference. Experts were invited as commentators to facilitate in-depth discussion. Six conferences were held between 2020 and 2022. After the fourth and sixth conferences, anonymous multiple-choice online surveys were administered to all attendees. Results: Participants included trainees and faculty. At each conference, 3 to 5 cases of rare endocrine diseases from up to 4 institutions were presented, mainly by trainees. Sixty-two percent of attendees reported 4 facilities as the appropriate size for the collaboration to maintain active learning in case conferences. Eighty-two percent of attendees preferred a semiannual conference. The survey also revealed the positive impact on trainees' learning regarding diversity of medical practice, academic career development, and confidence in honing of presentation skills. Conclusion: We present an example of our successful virtual global case conference to enhance learning about rare endocrine cases. For the success of the collaborative case conference, we suggest smaller cross-country institutional collaborations. Preferably, they would be international, semiannually based, and with recognized experts as commentators. Since our conference has engendered multiple positive effects on trainees and faculty, continuation of virtual education should be considered even after the pandemic era.

A Case of Hypophysitis Associated With SARS-CoV-2 Vaccination.

Background/Objective: Although SARS-CoV-2 vaccines have been developed with multiple novel technologies and rapidly disseminated worldwide, the full profile of adverse effects has not been known. Recently, there are sporadic but increasing reports of endocrinopathy in relation to SARS-CoV-2 vaccination. Here we report a rare case of hypophysitis with acute onset of diabetes insipidus, immediately after SARS-CoV-2 vaccination. Case Report: A 48-year-old female patient had been in her usual state of health until she received the first SARS-CoV-2 vaccine. Two days after vaccination, she started to have flu-like symptoms, including severe headache and myalgia as well as persistent headache, polydipsia, and polyuria. She was diagnosed with diabetes insipidus, and magnetic resonance imaging revealed thickening of the pituitary stalk. Three months after vaccination, her symptoms had somewhat improved, but she still had pituitary stalk thickening on magnetic resonance imaging. Discussion: Given the timing of the occurrence of diabetes insipidus, we believe that the patient's hypophysitis may be associated with SARS-CoV-2 vaccination. We also found 19 cases of endocrinopathy after SARS-CoV-2 vaccination by literature search. The reported endocrine organs were the thyroid, pituitary, and adrenals. Twelve cases of diabetes were also reported. Among 3 pituitary cases, diabetes insipidus was reported only in our case. Conclusion: We report a rare case of SARS-CoV-2 vaccine-triggered hypophysitis, which led to diabetes insipidus. SARS-CoV-2 vaccine-related endocrinopathy seems, indeed, possible. Endocrinopathy is associated with infrequent complications; however, it may be underestimated in the post-SARS-CoV-2-vaccinated population. Further studies are warranted to better understand SARS-CoV-2 vaccine-related endocrinopathy.

The degree of saturation of fatty acids in dietary fats does not affect the metabolic response to ingested carbohydrate.

BACKGROUND: We are interested in the metabolic response to ingested macronutrients, and the interaction between macronutrients in meals. Previously, we and others reported that the postprandial rise in serum glucose following ingestion of 50 g carbohydrate, consumed as potato, was markedly attenuated when butter was ingested with the carbohydrate, whereas the serum insulin response was little affected by the combination. OBJECTIVE: To determine whether a similar response would be observed with three other dietary fats considerably different in fatty acid composition. DESIGN: Nine healthy subjects received lard, twelve received olive oil and eleven received safflower oil as a test meal. The subjects ingested meals of 25 g fat (lard, olive oil or safflower oil), 50 g CHO (potato), 25 g fat with 50 g CHO or water only. Glucose, C peptide, insulin, triacylglycerols and nonesterified fatty acids were determined. RESULTS: Ingestion of lard, olive oil or safflower oil with potato did not affect the quantitative glucose and insulin responses to potato alone. However, the responses were delayed, diminished and prolonged. All three fats when ingested alone modestly increased the insulin concentration when compared to ingestion of water alone. When either lard, olive oil or safflower oil was ingested with the potato, there was an accelerated rise in triacylglycerols. This was most dramatic with safflower oil. CONCLUSIONS: Our data indicate that the glucose and insulin response to butter is unique when compared with the three other fat sources varying in their fatty acid composition.

State representation in mental illness.

Reinforcement learning theory provides a powerful set of computational ideas for modeling human learning and decision making. Reinforcement learning algorithms rely on state representations that enable efficient behavior by focusing only on aspects relevant to the task at hand. Forming such representations often requires selective attention to the sensory environment, and recalling memories of relevant past experiences. A striking range of psychiatric disorders, including bipolar disorder and schizophrenia, involve changes in these cognitive processes. We review and discuss evidence that these changes can be cast as altered state representation, with the goal of providing a useful transdiagnostic dimension along which mental disorders can be understood and compared.

Holistic Reinforcement Learning: The Role of Structure and Attention.

Compact representations of the environment allow humans to behave efficiently in a complex world. Reinforcement learning models capture many behavioral and neural effects but do not explain recent findings showing that structure in the environment influences learning. In parallel, Bayesian cognitive models predict how humans learn structured knowledge but do not have a clear neurobiological implementation. We propose an integration of these two model classes in which structured knowledge learned via approximate Bayesian inference acts as a source of selective attention. In turn, selective attention biases reinforcement learning towards relevant dimensions of the environment. An understanding of structure learning will help to resolve the fundamental challenge in decision science: explaining why people make the decisions they do.

Thyroid nodules over 4 cm do not have higher malignancy or benign cytology false-negative rates.

PURPOSE: Whether thyroid nodules 4 cm or larger with benign cytology carry a higher risk of malignancy, and should be managed differently than smaller nodules remains controversial. We aimed to evaluate the malignancy rate and benign cytology false-negative rate in thyroid nodules ≥4 cm compared with those <4 cm. METHODS: All thyroidectomies between January 2010 and December 2014 were reviewed. Patient demographics, preoperative sonographic nodule size, fine needle aspiration cytology (FNAC), and final surgical pathology results were compared for index nodules ≥4 vs. <4 cm. RESULTS: A total of 490 index nodules with preoperative FNAC were identified. A total of 137 nodules were ≥4 cm and 353 nodules were <4 cm. The prevalence of carcinoma was lower (23 vs. 53%) in nodules ≥4 vs. <4 cm (p < 0.0001). The false-negative rate of benign FNAC for ≥4 and <4 cm index nodule was 5.2% and 5.9%, respectively (p = 1.000). CONCLUSIONS: This study shows that thyroid nodules ≥4 cm do not have a higher malignancy rate at surgery nor higher benign cytology false-negative rate than smaller nodules. Thyroid nodules over 4 cm do not require resection, to rule out malignancy, based on size alone.

Dynamic Interaction between Reinforcement Learning and Attention in Multidimensional Environments.

Little is known about the relationship between attention and learning during decision making. Using eye tracking and multivariate pattern analysis of fMRI data, we measured participants' dimensional attention as they performed a trial-and-error learning task in which only one of three stimulus dimensions was relevant for reward at any given time. Analysis of participants' choices revealed that attention biased both value computation during choice and value update during learning. Value signals in the ventromedial prefrontal cortex and prediction errors in the striatum were similarly biased by attention. In turn, participants' focus of attention was dynamically modulated by ongoing learning. Attentional switches across dimensions correlated with activity in a frontoparietal attention network, which showed enhanced connectivity with the ventromedial prefrontal cortex between switches. Our results suggest a bidirectional interaction between attention and learning: attention constrains learning to relevant dimensions of the environment, while we learn what to attend to via trial and error.

The effects of aging on the interaction between reinforcement learning and attention.

Reinforcement learning (RL) in complex environments relies on selective attention to uncover those aspects of the environment that are most predictive of reward. Whereas previous work has focused on age-related changes in RL, it is not known whether older adults learn differently from younger adults when selective attention is required. In 2 experiments, we examined how aging affects the interaction between RL and selective attention. Younger and older adults performed a learning task in which only 1 stimulus dimension was relevant to predicting reward, and within it, 1 "target" feature was the most rewarding. Participants had to discover this target feature through trial and error. In Experiment 1, stimuli varied on 1 or 3 dimensions and participants received hints that revealed the target feature, the relevant dimension, or gave no information. Group-related differences in accuracy and RTs differed systematically as a function of the number of dimensions and the type of hint available. In Experiment 2 we used trial-by-trial computational modeling of the learning process to test for age-related differences in learning strategies. Behavior of both young and older adults was explained well by a reinforcement-learning model that uses selective attention to constrain learning. However, the model suggested that older adults restricted their learning to fewer features, employing more focused attention than younger adults. Furthermore, this difference in strategy predicted age-related deficits in accuracy. We discuss these results suggesting that a narrower filter of attention may reflect an adaptation to the reduced capabilities of the reinforcement learning system. (PsycINFO Database Record

DYT1 dystonia increases risk taking in humans.

It has been difficult to link synaptic modification to overt behavioral changes. Rodent models of DYT1 dystonia, a motor disorder caused by a single gene mutation, demonstrate increased long-term potentiation and decreased long-term depression in corticostriatal synapses. Computationally, such asymmetric learning predicts risk taking in probabilistic tasks. Here we demonstrate abnormal risk taking in DYT1 dystonia patients, which is correlated with disease severity, thereby supporting striatal plasticity in shaping choice behavior in humans.

Correlation Between Histological Diagnosis and Mutational Panel Testing of Thyroid Nodules: A Two-Year Institutional Experience.

BACKGROUND: Indeterminate thyroid fine-needle aspiration (FNA) cytology, including atypia of undetermined significance (AUS/FLUS) and suspicious for follicular neoplasm (SFN), continues to generate uncertainty about the presence of malignancy, resulting in repeated follow-up, repeat FNA, or diagnostic surgery. Mutational panel testing may improve the malignancy risk prediction in indeterminate nodules, but the general application of such testing has not been investigated extensively. METHODS: A retrospective review was performed of all patients undergoing thyroidectomy at a tertiary care facility over a two-year period. Mutational panel test results, when present, were analyzed relative to FNA cytologic result and surgical histopathologic diagnosis. Malignancy rates, sensitivity, specificity, positive predictive values (PPV), negative predictive values (NPV) and positive and negative likelihood ratios (LR) were calculated. RESULTS: A total of 261 operated thyroid nodules had the following initial FNA cytology results: 2% non-diagnostic, 23% benign, 28% AUS/FLUS, 11% SFN, 9% suspicious for malignancy (SUSP), and 27% malignant. The histopathologic malignancy rate was 48%, subcategorized by cytology into benign 7%, AUS/FLUS 30%, SFN 38%, and SUSP 83%. Mutations were more frequent in indeterminate nodules that were histologically malignant versus benign (p < 0.0001) or versus adenoma (p = 0.001). Mutational analysis in 44 AUS/FLUS nodules resulted in a malignancy detection sensitivity of 85%, a specificity of 65%, a PPV of 50%, a NPV of 91%, and a positive LR of 2.4. In 12 SFN nodules analyzed with ThyroSeq(®) testing, sensitivity was 100%, specificity 57%, PPV 63%, NPV 100%, and LR 2.3. Performance of the seven-gene mutational panel was not significantly different from the ThyroSeq(®) panel in the AUS/FLUS group. The malignancy yield, comparing the mutation positive AUS/FLUS group with the untested AUS/FLUS surgical cohort, did not reach statistical significance (p = 0.17). CONCLUSIONS: In a surgical cohort, a similar NPV but a lower PPV was found with the use of mutational panel testing compared to the published literature. Following the identification of a mutation, the prevalence of malignancy in the AUS/FLUS or SFN category was increased by nearly 15% to 45% and 53%, respectively. Further study is needed to confirm these results and to analyze clinical outcome subcategories relative to the utility of mutational testing.

A free-choice premium in the basal ganglia.

Apparently, the act of free choice confers value: when selecting between an item that you had previously chosen and an identical item that you had been forced to take, the former is often preferred. What could be the neural underpinnings of this free-choice bias in decision making? An elegant study recently published in Neuron suggests that enhanced reward learning in the basal ganglia may be the culprit.