114333 research outputs found
Sort by
Hybrid Cultivars and Straw Mulch Improve Yield and Gross Margin of Squash in Eastern Uganda
No full textSquash (Cucurbita maxima Duch. and Cucurbita moschata Duch.) remains an underutilized vegetable crop in Uganda, despite its nutritional and potential economic benefits. Squash is grown at a subsistence level with minimal input, and unimproved cultivars and prolonged drought challenge its production. Our study aimed to evaluate the impact of cultivar selection and soil mulching with straw on squash yield, marketability, and gross margin across 2 years: Oct 2020 to Feb 2021 (year 1) and Oct 2021 to Feb 2022 (year 2). Field trials were conducted in two subcounties of the Kamuli District, Butansi and Namasagali, in Eastern Uganda, in a randomized complete block design with a split-plot arrangement of treatments and replicated four times in each location and year. The main plot included soil mulching (+/−) using dried sugarcane (Saccharum officinarum L.) straw, and subplots included four cultivars: two hybrids, ‘Arjuna F1’ and ‘Pujinta F1’ (C. moschata), and two open-pollinated, ‘Orange Flesh’ and ‘Mayford’ (C. maxima). Hybrid cultivars had the highest total yields, marketable fruit numbers, and weights compared with open-pollinated cultivars during both years in Namasagali subcounty field plots and in Butansi subcounty plot in year 2. Mulched plots had the highest total and marketable fruit numbers and weights in both locations during year 2 and in Butansi subcounty in year 1. Gross margins were higher in mulched plots in years 1 and 2 in Namasagali and in year 1 in Butansi subcounties. Mulched plots had lower aboveground weed biomass and soil temperatures than nonmulched plots. The adoption of hybrid squash cultivars combined with straw mulch presents a viable strategy for enhancing the sustainability of agricultural production in Eastern Uganda. This approach demonstrates significant potential for enhancing yield, improving fruit marketability, and increasing gross margin.This article is published as Kwikiiriza, S., Nonnecke, G. R., Nair, A., Akitwine, F., & Burras, C. L. (2026). Hybrid Cultivars and Straw Mulch Improve Yield and Gross Margin of Squash in Eastern Uganda. HortTechnology, 36(1), 114–123. https://doi.org/10.21273/HORTTECH05798-2
Small-Form-Factor Multiple-Output Isolated DC-DC Converter Design
No full textGalvanically isolated DC-DC converters are essential in a variety of modern electronic systems, including industrial, automotive, grid-tied, and medical applications. They provide galvanic isolation between high-voltage (HV) and low-voltage (LV) domains, a critical function for protecting human beings and sensitive low-voltage circuitry from hazardous high voltages and destructive surge events.
Compared with traditional designs with discrete circuit, A key challenge in designing these converters, particularly miniature inductive isolated DC-DC converters that use small-form-factor micro-transformers, is balancing efficiency, integration, and performance. While emerging in-package/on-chip solutions have addressed the form factor, their efficiency has historically been low, peaking at 7-53% due to the low-quality factor of integrated transformers. The efficiency degradation is also caused by higher parasitic resistance, which leads to conduction loss, and the need to operate at higher frequencies, which increases switching loss.
Another significant challenge is implementing a fast and reliable closed-loop control system across the isolation barrier. Traditional designs often use linear control schemes like Pulse Width Modulation (PWM) or Pulse Frequency Modulation (PFM). These methods typically rely on digital isolators as feedback channels to close the isolated control loop, which tend to be bulky and increase system complexity and cost. This reliance on linear control also often leads to trade-offs between stability, dynamic response, and robustness against process, voltage, and temperature (PVT) variations.
To address these limitations, non-linear control techniques and novel architectures are being explored. The ultimate goal is to achieve reliable closed-loop control without the need for a dedicated data link across the isolation barrier. This research aims to overcome these challenges by developing innovative control schemes that leverage the power link itself for feedback, enabling fast and stable operation while maintaining a compact form factor.
This dissertation presents the design, analysis, and implementation of two small-form-factor isolated DC-DC converters for multiple-output applications. The work begins by introducing a Single-Link Multiple-Output (SLiMO) converter, which provides two regulated voltage rails with a single low-cost standard flexible printed circuits (FPC)-based micro-transformer design with 6-mm diameter and 0.13-mm thickness, achieving both local voltage regulation at the secondary side and global power regulation across the isolated barrier at the primary side. A digital discrete frequency modulation (DFM) is introduced as the control scheme for the global power regulation, and a current-rotation-based hysteretic control is introduced as the control scheme for the local voltage regulation of the active rectifier for the two outputs. The proposed SLiMO converter is fabricated in 0.18-μm CMOS using 3.3-V IO devices, achieving a 63% peak efficiency, a 1.29-W maximum output power, and a fast load- transient-response.
Although successfully , the work then introduces a novel Single-Link Multi-Domain-Output (SLiMDO) converter designed to provide two regulated outputs in separate domains in the receiver (Rx) side with a single micro- transformer. These isolated Rxes achieve local voltage regulation and automatic energy distribution across domains through passive magnetic flux sharing without any shared controllers or communication links for coordination. Unlike the SLiMO, the transmitter (Tx) achieves global power modulation by sensing Rx behaviors through the reflected impedance without using dedicated digital isolators as data links, thus enables closed-loop control with a fast transient response. The SLiMDO converter prototype consists of one Tx chip and two identical Rx chips for two outputs, all fabricated in 0.18-μm CMOS using 3.3-V I/O devices. The micro- transformer is fabricated in standard flexible-printed-circuit (FPC) with 8-mm diameter and 0.13-mm thickness. The isolation level of the transformer is measured to be >5kV between Tx and Rx, and >3kV between Rx and Rx. The SLiMDO operation is verified in measurements converting a 3.3-V input from the primary domain to two outputs from 1.8 to 3.3V each located in separated isolated secondary domains, with a 62.6% peak efficiency at ~600 mW, a 1.13-W maximum power, and a decent load transient response
Epigenetic mechanisms controlling gene expression in livestock immune cells
No full textUnderstanding the molecular mechanisms underlying immune cell regulation and innate immune memory is crucial for developing strategies to enhance disease resilience in both livestock and human populations. This dissertation presents a comprehensive multi-omics investigation of immune cell regulatory landscapes and trained immunity mechanisms across species, employing cutting-edge genomic approaches to elucidate the epigenetic and transcriptional networks governing innate immune cell function and memory formation.
This work comprises three interconnected research chapters that collectively establish a foundational framework for understanding immune cell regulation and trained immunity. In research described in chapter 2, I present the first single-nucleus chromatin accessibility sequencing (snATAC-seq) analysis of porcine peripheral blood mononuclear cells (PBMCs), identifying 17,207 nuclei across 35 clusters representing 12 distinct immune cell types. This study established a comprehensive atlas of 110,444 high-quality accessible chromatin regions and 11,872 unique cell type-specific differentially accessible peaks (DAPs), demonstrating concordant cell type annotations between snATAC-seq and scRNA-seq approaches. I also identified 244 cis-co-accessibility networks (CCANs), including 14 monocyte-specific networks associated with key immune genes (NLRP3, CSF1R, S100A8, GSDMD, and others), identified putative enhancer or promoter regions associated with differentially expressed genes (DEG) in each immune cell type including monocytes, and provided critical baseline regulatory frameworks for subsequent immunity studies.
In work described in chapter 3, I characterized BCG (Bacillus Calmette-Guérin)-induced trained immunity in porcine monocytes through integrative analysis of cytokine production, RNA-seq, and ATAC-seq. Neonatal BCG administration resulted in significant elevation of TNF protein expression in monocytes upon in vitro LPS stimulation (p=0.05) and dramatically increased the breadth of stimulation-responsive genes. BCG treatment dramatically increased the number of genes that responded to LPS stimulation: at 3 weeks post-BCG, we detected 2,998 LPS-responsive DEGs in BCG-treated pigs compared to only 525 in control pigs, and this effect persisted at 6 weeks (4,447 vs 2,917 DEGs, respectively). Among all LPS-responsive DEGs regardless of treatment group at week 6, over 3,150 DEGs showed altered magnitude of LPS responsiveness. Genome-wide chromatin accessibility remodeling due to BCG was evidenced by 4,805 differentially accessible peaks (3,396 increased, 1,409 decreased) in monocytes collected from BCG-administered compared to those in control pigs. Interestingly,133 genes met published criteria for "epigenetic potential", showing coordinated changes in both chromatin accessibility and stimulus responsiveness. Gene Ontology (GO) analysis of these 133 genes revealed significant enrichment in pathways highly relevant to trained immunity, including "innate immune response," "positive regulation of cytokine production," and “NF-κB signaling”, with remarkable conservation compared to human trained immunity studies.
Research described in chapter 4 demonstrated BCG-induced trained immunity mechanisms in bovine immune cells through integrative analysis of data measuring cytokine production, RNA-seq, as well as H3K27ac histone modification, an epigenetic marker for promoters and enhancers. BCG treatment significantly enhanced IL-1β and IL-6 production in response to subsequent LPS (p=0.02 and p=0.015) and Poly(I:C)/Imiquimod (p<0.0001) stimulation of monocytes in vitro, while increasing the breadth of stimulation-responsive genes (1,439 vs 1,214 DEGs for LPS vs. control; 6,513 vs 5,605 DEGs for Poly(I:C)/Imiquimod in BCG vs noBCG groups). Genome-wide H3K27ac profiling identified 227 differential peaks in monocytes and 383 in alveolar macrophages isolated from BCG treated compared to those from control calves, indicating BCG-induced epigenetic reprogramming across multiple immune cell types. GO analysis revealed selective enrichment of innate immune response pathways in BCG-enhanced gene lists but not in lists of control-enhanced genes.
Comparison of results from the above projects revealed important connections between baseline regulatory networks and BCG-induced modifications. The 14 monocyte CCANs identified in Chapter 2 were associated with genes highly relevant to trained immunity mechanisms, including inflammasome components and damage-associated molecular patterns. Several transcription factors identified as important for cell type identity or CCAN in Chapter 2 were also predicted as mediators of BCG-induced changes in Chapter 3. Notably, binding motifs of PU.1 and PU.1-IRF8 were enriched in both monocyte-specific DAPs in Chapter 2 and BCG-decreased peaks in Chapter 3, while PU.1-IRF binding motifs were only enriched in monocyte DAPs in Chapter 2, possibly reflecting the selective mechanisms of BCG-induced trained immunity. Cross-species pathway comparison revealed remarkable similarity in functional programs enhanced by trained immunity between porcine and bovine, with both species showing consistent enrichment in innate immune response, cytokine-mediated signaling, and inflammatory response pathways.
Chapter 3 and Chapter 4 work faced challenges related to substantial within-group variance (variance among biological replicates within each treatment-stimulation subgroup) across all omics platforms, attributed to genetic heterogeneity in outbred animal populations, unequal environmental exposures, and individual variation in training capacity. Additionally, tissue-resident alveolar macrophage analysis revealed methodological challenges related to cell population heterogeneity, highlighting the need for improved isolation protocols in future studies.
The comprehensive catalog of porcine immune cell regulatory elements contributes valuable resources to the Functional Annotation of Animal Genomes (FAANG) initiative for interpreting genetic variation in immune-related traits. Additionally, the findings provide mechanistic insights into how BCG induces long-term functional reprogramming of innate immune cells and demonstrate that fundamental mechanisms of trained immunity are evolutionarily conserved across mammalian species. These results have important implications for veterinary medicine, offering fundamental background for developing strategies to improve livestock disease resilience while reducing antibiotic dependence, providing insights into vaccination optimization and broad-spectrum immune enhancement approaches
Brains and bushels: Bioinformatic insights into Alzheimer’s risk and Maize evolution - computational approaches to cognitive decline and plant adaptation
No full textThis dissertation covers two topics: genetic risk factors of Alzheimer’s disease (AD) and evolution of environmental adaptation in maize.
The first topic involves research on AD, which is considered the most common cause of dementia in older adults aged 65 years and older and is the 7th most common cause of death in older adults. Carrying the Apolipoprotein E ε4 (APOE4) isoform emerged decades ago as a risk factor for the development of AD, leading to cognitive effects that include difficulties with both executive function and memory. The literature involving the effect of the APOE isoforms on cognitive function is inconsistent. However, many of these studies categorized as APOE as a binary variable with participants either being “carriers” of at least one copy of the APOE4 isoform or “non-carriers” having no copies of the APOE4 isoform. The protective isoform has been the subject of fewer studies and no studies to my knowledge have compared results from using different methods to categorize APOE.
A finer categorization of APOE isoforms (protective, neutral, risk and mixed) indicates that the previous APOE carrier classification was driven primarily by a subset of the isoforms, specifically the APOE4 isoform and not the mixed APOE4/APOE2 mixed isoform. We demonstrate the importance of careful categorization when studying allelic effects and their interaction with covariates in the model.
APOE is not the only proposed genetic risk factor for AD. Several SNPs on the adjacent TOMM40 gene have also been identified as potential genetic risk factors. When TOMM40 rs2075650 SNP is included in the model, we uncover a stronger individual APOE effect and characterize the independent and interactive effects of TOMM40 and APOE and covariates of sex, age, education and socioeconomic status on cognitive decline.
The second topic involves environmental adaptation in maize after its domestication. we sought to further understand the ability of maize to adapt to environments encompassing nearly the entire world to become one of the most abundant agricultural crops. Specifically, we worked to answer the following research questions: 1) How does the pan-gene classification and age of the gene affect tissue specificity measures; 2) Are there consistently differentially expressed pan-genes across multiple tissues between tropical and temperate varieties of maize; and 3) Does the network structure among pan-genes change between tropical and temperate varieties. To address the first question, we used the dataset from (Hufford et al., 2021) and tissues specificity measures of tau and coefficient of variation (COV). The second research question is addressed using a modified differential expression pipeline that allows for preserving a unique aspect of this data in that two types of missing data (NAs and zeros) exist in the data with different indications important for interpretation. Finally, the third question is addressed using weighted gene co-expression network analysis (WGCNA) to compare pan-gene networks constructed for tropical and temperate lines of Maize
A qualitative study on the implementation of best practices for safeguarding conveyors in industrial facilities in Iowa
No full textThis qualitative study, using a grounded theory approach, examined best practices for safeguarding conveyors in industrial facilities across Iowa, focusing on hazard identification, risk assessment, and employee training. Semi-structured interviews were conducted with five Certified Safety Professionals (CSPs) from various industries, each with 13 to 21 years of experience. Key findings included effective safeguarding strategies such as trained in-house auditors, cross-functional assessment teams, and customized hazard evaluations for routine and non-routine operations. Participants emphasized the importance of incorporating industry safety standards, developing alternative procedures, and engaging employees through targeted training and mentoring. In conclusion, the study demonstrated that proactive safety management systems, technical expertise, collaborative hazard assessments, and team engagement are key aspects in reducing injuries related to conveyors
Discovery of novel sources of resistance to maize lethal necrosis (MLN)
No full textMaize lethal necrosis (MLN), caused by the synergistic infection of maize chlorotic mottle virus (MCMV) and sugarcane mosaic virus (SCMV), poses a major threat to maize production in Sub-Saharan Africa (SSA). Currently, two resistance sources are available: the qMLN.06_157 allele, which confers resistance to MCMV, and edited eIF4E1/eIF4E2 alleles, which provide resistance to SCMV. Individually, these traits result in full MLN resistance. However, the rapid evolution of viruses demands the need to broaden the sources of resistance for durable control. This study aimed to identify novel genetic bases of MLN resistance using three complementary approaches: transcriptomics, genome editing, and comparative genomics.
RNA-seq analysis of resistant and susceptible maize genotypes revealed a large number of differentially expressed genes (DEGs). Many of these were associated with viral resistance pathways that are well known to play critical roles in plant–virus interactions. These included innate immune signaling and redox regulation, which are central to basal defense; RNA interference (RNAi), a key antiviral mechanism against potyviruses such as SCMV; the ubiquitin–proteasome system, which mediates protein degradation during infection; and DNA damage repair pathways that maintain genome integrity under stress. Although not the most statistically significant, subtle yet consistent shifts were also observed in the expression of translation initiation factors, hinting at their potential role in modulating host–virus interactions during MLN infection.
To directly evaluate the involvement of translation-related mechanisms, CRISPR-Cas9 editing was employed to disrupt four eIF4G and eIF(iso)4G genes in the fast-flowering maize line minimaize. These genes were selected due to their established role in virus resistance in other cereals, such as rice. Despite successful edits, all plants remained susceptible to SCMV, suggesting that either eIF4Gs are not essential for SCMV infection in maize or that functional redundancy among gene family members masked the effects of individual mutations. While resistance was not achieved, the resulting edited lines provide a valuable resource for further dissecting the recruitment of host translation factors by viruses, especially potyviruses.
Complementing these molecular approaches, comparative genomic analyses of previously reported MLN-resistance QTL on chromosomes 1, 3, 5, 6, and 9 uncovered additional candidate genes and variants that differentiated resistant from susceptible genotypes. Many of these genes map to pathways linked to viral replication, RNA silencing, translation, vesicle trafficking, and viral movement, providing new leads for functional validation.
Together, these approaches expand the pool of candidate genes and pathways, offering important opportunities for biotechnological intervention and breeding against MLN
Ion sensing in agricultural and food applications: a critical review of solid-contact ion-selective sensors
No full textSolid-contact ion-selective electrodes (SC-ISEs) have gained prominence as versatile tools for monitoring ionic species in complex agricultural and food matrices, offering low-cost, miniaturizable, and field-deployable alternatives to conventional laboratory assays. This review critically examines electrochemical transduction techniques for SC-ISEs, including potentiometry, coulometry, amperometry, voltammetry, and electrochemical impedance spectroscopy. By evaluating these techniques based on their detection mechanisms, sensitivity ranges, selectivity characteristics, and application potential, we highlight how dynamic electrochemical modes can overcome potentiometry limitations. Emphasis is placed on the role of solid-contact design, nanostructured materials (e.g., conducting polymers, carbon nanomaterials, and laser-induced graphene), and integrated readout strategies that enhance sensor performance in real-world applications. We also analyze state-of-the-art configurations for ion detection in soil, water, and food products. Finally, we discuss current challenges and offer perspectives on SC-ISEs integration into cyber-physical systems, where real-time, connected, and autonomous sensing will be central for sustainable agriculture and food systems, while also addressing regulatory considerations.This is a manuscript of an article published as Miliao, G.L., Winandar, F.G., Leung, E.H. et al. Ion sensing in agricultural and food applications: a critical review of solid-contact ion-selective sensors. Anal Bioanal Chem (2026).doi: https://doi.org/10.1007/s00216-025-06302-3.We thankfully acknowledge funding support from the National Institute of Food and Agriculture, U.S. Department of Agriculture, award numbers 2020-67021-31375, 2021-67021-34457, 2021- 67011-35130, 2023-38821-39796, and 2018-67016-27578, awarded as a Center of Excellence, the National Science Foundation under award numbers CMMI-2037026, EEC-2231632, and 2231632 awarded as a IUCRC Phase I center in addition to the Digital and Precision Agriculture Applications Funding Opportunity at the Iowa State University
Beyond single plots: A benchmark for question answering on scientific multi-charts
No full textCharts are widely used in many fields, where interpreting them correctly is key to informed decisions. In many real-world scenarios, a single chart is insufficient; instead, multiple relevant charts must be understood together to extract meaningful insight and make decisions. Research on understanding multi-chart images has not been extensively explored. In this work, we introduce PolyChartQA, a large-scale dataset specifically designed for question answering over multi-chart figures. PolyChartQA comprises 534 multi-chart images (with a total of 2,297 sub-charts) sourced from peer-reviewed computer science research publications and 2,565 question-answer pairs annotated across diverse question types and difficulty levels. We evaluate the performance of eight state-of-the-art Multimodal Language Models (MLMs) on PolyChartQA across question type, difficulty, question source, and key structural characteristics of multi-charts. Our findings reveal the following insights. (1) Our proposed prompting method improves L accuracy (LLM-based accuracy) up to 11.03% for multi-chart question answering. (2) Significant gaps remain for MLMs to understand multi-charts correctly, as evidenced by a 27.7% L-accuracy drop on human-authored vs. MLM-generated questions
Decoding the catalytic plasticity in class II diterpene cyclases
No full textDiterpenoids form a subclass of terpenoids that are made of twenty carbon atoms, with over 35,000 identified natural products. These molecules are widely used in industry as additives in food, fragrance, and cosmetic products, as well as in pharmaceuticals for their antimicrobial, anti-inflammatory, and anticancer activities. The labdane-related diterpenoids form a superfamily of >15,000 compounds, defined by use of a Class II diterpene cyclase (DTCs). For example, these are required to initiate gibberellin (GA) phytohormones biosynthesis and a broad range of more specialized secondary metabolites. Altered catalytic base residues in DTCs have been shown to be sufficient to change the product outcome. For instance, a catalytic base dyad formed by histidine and asparagine residues from the LHS and PNV motifs produces ent-labda-8(17),13-dien-15-yl diphosphate (ent-CPP) for GA biosynthesis in ancestral DTCs. Surprisingly, substituting either residue of this dyad altered the product. For instance, replacing histidine with tyrosine yields a fully rearranged product, ent-kolavenyl diphosphate (KPP), while substitution of either residue with alanine results in water incorporation, producing 8α-hydroxy-ent-labda-13-en-15-yl diphosphate (LPP). We investigated the role of these residues through ancestral residue restoration in three DTCs from non-seed plants; kolavenol synthase from the liverwort Odontoschisma prostratum (OpKOS), isoabienol synthase from the moss Orthotrichum lyellii (OlIAS), and isopimara-8,15-diene synthase from the fern Vittaria appalachiana (VaIPS), highlighting the evolutionary use of residue switches for product diversification. Moreover, the abietaenol synthase from Abies grandis (AgAS), a model for the TPS-d3 sub-family involved in conifer resin acid biosynthesis, has been hypothesized to employ a hydrogen-bonded histidine–tyrosine pair in its DTC active site as the catalytic base. Substitution of histidine with aspartate or tyrosine with phenylalanine results in water incorporation, leading to production of the hydroxylated CPP derivative labda-13-en-8β-ol-15-yl diphosphate (LPP) and a double bond isomer of CPP (i.e., 7-endo CPP). However, the precise identity of the native catalytic base, as well as the alternative functional group responsible for the formation of the formation of LPP and 7-endo-CPP upon such substitutions, remains unresolved. We applied the TerDockin computational approach, combining quantum chemical modeling with computational docking, to the AgAS DTC active site to decode the catalytic bases for the observed product profile and verified their identity by mutational analysis. Additionally, structural analysis of copalyl diphosphate synthase (CPS) from Arabidopsis thaliana (AtCPS), together with sequence comparisons from various DTCs, identified seven aromatic residues in the active site that are highly conserved, suggesting essential roles in stabilizing carbocation intermediates through cation–π interactions. To assess their functional importance in determining product specificity, each aromatic residue was systematically substituted in two representative CPS enzymes: AtCPS from plants and EtCPS from the bacterium Erwinia tracheiphila. Overall, the work presented in this dissertation increases our mechanistic understanding of the catalytic plasticity of DTCs, which underlies the observed diversification of the derived labdane-related diterpenoid natural products (>15,000 known)
Polarized risk shift in an operational domain: decision-making approaches for individuals versus teams
No full textThis study sought to find a polarized risk shift between individual risk decisions and team risk decisions in an operational task. Risk involves the potential for loss, reward, and uncertainty, with both individuals and teams routinely encountering risk decisions. In high-stakes operational domains such as military command, aviation, and emergency response, decision-makers act under incomplete, ambiguous, or competing information, where failures in team cognition and communication can compromise safety, performance, and decision quality. Studies have shown that teams often exhibit polarization when making risk-related decisions.
Risk polarization has been explored in behavioral psychology and teamwork literature and is defined as the propensity of teams to make more extreme risk decisions when compared to individuals. However, prior research on polarized decision-making has largely relied on abstract, hypothetical tasks that lack relevance to operational contexts, limiting the applicability of findings to real-world environments. These previous tasks lacked attributes present in an operational domain, including complexity, accountability, realism, and measurable risk.
This study aims to bridge this gap by developing an applied, operational task that mirrors real-world scenarios, enabling investigation of whether this polarization occurs in a high-stakes, operational domain. Humans often work in teams in these domains, and increasingly, humans are also working with autonomous agents.
An evaluation was done to first establish the existence of polarization in human-only teams in an operational setting. The study had one independent variable: decision-maker with two levels, individual and team. The task design involved an aviation dispatch task, where participants simulated the role of a flight dispatcher. Participants were responsible for making dispatch decisions on whether to divert, hold, or send 25 airplanes, while weighing the potential consequences of incurring policy violations based on the possibility of an approaching storm.
Teams demonstrated a significant polarized risk shift with decision shifts moving bidirectionally, where roughly half of the teams became riskier and half more cautious in their team decisions. Confidence significantly increased, and perceived responsibility significantly decreased in the team trial. Additionally, thematic analysis found that both informational and normative sources of influence contributed to the team decision-making process with three distinct decision strategies emerging in the team trial. These decision strategies help explain the directionality of the shifts, with some strategies promoting consensus towards riskier choices and others reinforcing caution and risk-aversion.
This study advances understanding of team risk decision-making by demonstrating significant polarized risk shift in an operational task that includes elements of complexity, accountability, realism, and measurable risk that were absent from prior research. These findings confirm that teams make more extreme decisions than individuals, and the thematic analyses conducted in this study on team decision dynamics and strategies indicate that this polarization emerges from team composition and the interplay of informational and normative influence processes. The insights from these thematic analyses provide design recommendations for developing effective team interventions and set the foundation for future experiments to examine how humans and autonomous agents perform teaming tasks in dynamic contexts involving risk. Future research should explore how agent design and human-autonomy team composition can be optimized to enhance decision quality and mitigate polarization