Expert Performance Research Articles

Uncertainty-aware automatic TNM staging classification for [18F] Fluorodeoxyglucose PET-CT reports for lung cancer utilising transformer-based language models and multi-task learning

Background[18F] Fluorodeoxyglucose (FDG) PET-CT is a clinical imaging modality widely used in diagnosing and staging lung cancer. The clinical findings of PET-CT studies are contained within free text reports, which can currently only be categorised by experts manually reading them. Pre-trained transformer-based language models (PLMs) have shown success in extracting complex linguistic features from text. Accordingly, we developed a multi-task ‘TNMu’ classifier to classify the presence/absence of tumour, node, metastasis (‘TNM’) findings (as defined by The Eight Edition of TNM Staging for Lung Cancer). This is combined with an uncertainty classification task (‘u’) to account for studies with ambiguous TNM status.Methods2498 reports were annotated by a nuclear medicine physician and split into train, validation, and test datasets. For additional evaluation an external dataset (n = 461 reports) was created, and annotated by two nuclear medicine physicians with agreement reached on all examples. We trained and evaluated eleven publicly available PLMs to determine which is most effective for PET-CT reports, and compared multi-task, single task and traditional machine learning approaches.ResultsWe find that a multi-task approach with GatorTron as PLM achieves the best performance, with an overall accuracy (all four tasks correct) of 84% and a Hamming loss of 0.05 on the internal test dataset, and 79% and 0.07 on the external test dataset. Performance on the individual TNM tasks approached expert performance with macro average F1 scores of 0.91, 0.95 and 0.90 respectively on external data. For uncertainty an F1 of 0.77 is achieved.ConclusionsOur ‘TNMu’ classifier successfully extracts TNM staging information from internal and external PET-CT reports. We concluded that multi-task approaches result in the best performance, and better computational efficiency over single task PLM approaches. We believe these models can improve PET-CT services by assisting in auditing, creating research cohorts, and developing decision support systems. Our approach to handling uncertainty represents a novel first step but has room for further refinement.

Evaluation of Vestibular Schwannoma Size Over Time: How Well do the Experts Perform and What can be Improved?

2D linear measurements are often used in routine clinical practice during vestibular schwannoma (VS) follow-up, primarily due to wider availability and ease of use. We sought to determine radiologist performance compared to 3D-volumetry, along with the impact of number of linear measurements, slice thickness and tumor volumes on these parameters. Single center retrospective study with 97 patients (592 MRI studies). Overall agreement, sensitivity, specificity and accuracy estimates and 95% confidence intervals were calculated for the entire cohort, and subgroups based on volumes (<400, 400-800, >800 mm3), slice thickness (≤ 1.5 mm or > 1.5 mm) and number of linear dimensions measured in the radiology report (0-1 or 2-3). There was weak agreement between radiologist inference and VS volumetry (0.45, 95% CI [0.41, 0.53]). Agreement was lower when 0-1 tumor dimension was measured (0.29, 95% CI [0.21, 0.42]), for smaller tumors < 400 mm3 (0.37, 95% CI [0.28, 0.45]), and for thick section imaging > 1.5 mm (0.36, 95% CI [0.25, 0.46]). The reader sensitivity was modest (0.49-0.54), while the accuracy for detecting ≤ ±25% interval change was weak (0.32-0.38). Reader performance trended towards improvement with thin-section imaging, measurement of 2-3 VS dimensions and for larger tumors. In routine practice, radiologists show poor agreement with volumetric results and sensitivity to detect interval change and overall poor accuracy for volumetric changes ≤ ± 25% in volume. In the absence of volumetric measurements, radiologists need to be more diligent when evaluating for interval changes in VS.

Large language models in radiology: Fluctuating performance and decreasing discordance over time

ObjectiveSince the introduction of large language models (LLMs), near expert level performance in medical specialties such as radiology has been demonstrated. However, there is limited to no comparative information of model performance, accuracy, and reliability over time in these medical specialty domains. This study aims to evaluate and monitor the performance and internal reliability of LLMs in radiology over a three-month period. MethodsLLMs (GPT-4, GPT-3.5, Claude, and Google Bard) were queried monthly from November 2023 to January 2024, utilizing ACR Diagnostic in Training Exam (DXIT) practice questions. Model overall accuracy and by subspecialty category was assessed over time. Internal consistency was evaluated through answer mismatch or intra-model discordance between trials. ResultsGPT-4 had the highest accuracy (78 ± 4.1 %), followed by Google Bard (73 ± 2.9 %), Claude (71 ± 1.5 %), and GPT-3.5 (63 ± 6.9 %). GPT-4 performed significantly better than GPT-3.5 (p = 0.031). Over time, GPT-4′s accuracy trended down (82 % to 74 %), while Claude’s accuracy increased (70 % to 73 %). Intra-model discordance rates decreased for all models, indicating improved response consistency. Performance varied by subspecialty, with significant differences in the Chest, Physics, Ultrasound, and Pediatrics sections. Models struggled with questions requiring detailed factual knowledge but performed better on broader interpretive questions. ConclusionLLMs, except GPT-3.5, performed above 70%, demonstrating substantial subject-specific knowledge. However, performance fluctuated over time, underscoring the need for continuous, radiology-specific standardized benchmarking metrics to gauge LLM reliability before clinical use. This study provides a foundational benchmark for future LLM performance evaluations in radiology.

Perform “like a king”: Novel learning strategies to optimise music students’ concert stage mastery

The role of tertiary music performance education is to prepare students for professional performance. Professional performers demonstrate stage mastery, and their concert preparation transcends the score to equip them for the concert stage. Music educators must translate professionals’ insights into performance education. This study developed, trialled and evaluated two learning modules for performance students, based on expert strategies. In Module 1, students engaged in a backstage immersion task. In Module 2, students roleplayed as expert performers and audience members. Participants reflected on their experiences in interviews. Workshops and interviews were recorded, transcribed, and coded thematically. Students analysed the concert environment critically and observed their performance mindset. Emulating expert performers prompted compelling stage presence and students observed the effects of stagecraft on audiences. Peer-to-peer learning facilitated critical reflection on professional practice and students resolved to apply these skills to concert preparation. Future studies will develop these tools for pedagogues and students.

Matching patients to clinical trials with large language models

Patient recruitment is challenging for clinical trials. We introduce TrialGPT, an end-to-end framework for zero-shot patient-to-trial matching with large language models. TrialGPT comprises three modules: it first performs large-scale filtering to retrieve candidate trials (TrialGPT-Retrieval); then predicts criterion-level patient eligibility (TrialGPT-Matching); and finally generates trial-level scores (TrialGPT-Ranking). We evaluate TrialGPT on three cohorts of 183 synthetic patients with over 75,000 trial annotations. TrialGPT-Retrieval can recall over 90% of relevant trials using less than 6% of the initial collection. Manual evaluations on 1015 patient-criterion pairs show that TrialGPT-Matching achieves an accuracy of 87.3% with faithful explanations, close to the expert performance. The TrialGPT-Ranking scores are highly correlated with human judgments and outperform the best-competing models by 43.8% in ranking and excluding trials. Furthermore, our user study reveals that TrialGPT can reduce the screening time by 42.6% in patient recruitment. Overall, these results have demonstrated promising opportunities for patient-to-trial matching with TrialGPT.

Role of Posterior Medial Thalamus in the Modulation of Striatal Circuitry and Choice Behavior.

The posterior medial (POm) thalamus is heavily interconnected with sensory and motor circuitry and is likely involved in behavioral modulation and sensorimotor integration. POm provides axonal projections to the dorsal striatum, a hotspot of sensorimotor processing, yet the role of POm-striatal projections has remained undetermined. Using optogenetics with slice electrophysiology, we found that POm provides robust synaptic input to direct and indirect pathway striatal spiny projection neurons (D1- and D2-SPNs, respectively) and parvalbumin-expressing fast spiking interneurons (PVs). During the performance of a whisker-based tactile discrimination task, POm-striatal projections displayed learning-related activation correlating with anticipatory, but not reward-related, pupil dilation. Inhibition of POm-striatal axons across learning caused slower reaction times and an increase in the number of training sessions for expert performance. Our data indicate that POm-striatal inputs provide a behaviorally relevant arousal-related signal, which may prime striatal circuitry for efficient integration of subsequent choice-related inputs.

Advocating individual-based profiles of elite athletes to capture the multifactorial nature of elite sports performance

Elite athletes are high-performance outliers within their specific sports. Even though science seeks to understand the nature of expertise and elite performance, much knowledge remains compartmentalized within subdisciplines. Despite this multidimensionality being acknowledged, an interdisciplinary approach to understanding elite athletes is still rare. This paper synthesizes insights across scientific domains in order to describe the population and individual characteristics of elite athletes. We analyzed diagnostic data from approximately 300 German squad athletes across eight different sports (e.g., gymnastics, volleyball, ice hockey, 3 × 3 basketball etc., agefemale = 18.95 ± 4.84 years, agemale = 19.32 ± 4.19 years) with expertise values ranging from 2 (low expertise) to 16 (high expertise). Data covered muscular strength, lower-body dynamics, muscle-power genetics, blood micronutrients, basic cognitive function, mental health, social support, and training conditions. Results of logistic regressions identified basic cognitive function (B = 0.89) and well-balanced blood micronutrients (B = 1.22) as critical factors distinguishing elite athletes. Additionally, multiple linear regressions suggested that lower-body dynamics (ß = 0.72) is related to increasing expertise values. We examined interactions between determinants of elite performance, and found that social support is positively associated with mental health and training conditions, whereas muscular strength correlates with lower-body dynamics. Focusing on top elite athletes in contrast to semi-elite athletes, we found higher within-group similarities in basic cognitive function and blood micronutrients. Findings indicate the need for a systemic, individualized, and comprehensive model using individual-based profiles.

Influence of Teachers’ Experience in Information Communication Technology and Students’ Academic Performance in Mathematics in Public Secondary Schools of Rwanda: A Case of Kayonza District

The study evaluated the impact of teachers' ICT experience on students' math performance in public secondary schools in Kayonza District, Rwanda. Involving 259 participants (223 teachers and 36 head teachers), data were gathered through questionnaires, interviews, and observations, and analyzed using SPSS Version 20.0. Findings revealed that 83.5% of respondents strongly agreed that teachers skilled in MS Excel, 85.9% in PowerPoint, and 83.9% in preparing teaching materials with ICT tools positively influenced students' performance. Additionally, 82.3% noted that teachers trained in exam preparation with ICT tools demonstrated their ICT experience. Improved student outcomes were reported, with 79.4% noting better numeracy, 80.2% improved grades, and 78.3% enhanced skills in experimenting with MS Excel. A significant positive correlation was found between teachers' ICT expertise and students' math performance. The study suggests that proficiency in ICT tools enhances students' academic performance. The Ministry of Education should support ICT integration to boost academic achievement, with superintendents and head teachers optimizing ICT training and lab usage. Teachers should share ICT skills to improve instruction and student outcomes. Future research should investigate the broader effects of ICT in education on academic achievement in mathematics.

Development of an Artificial Intelligence System for Distinguishing Malignant from Benign Soft Tissue Tumors Using Contrast-Enhanced MR Images.

The integration of artificial intelligence (AI) into orthopedics has enhanced the diagnosis of various conditions; however, its use in diagnosing soft-tissue tumors remains limited owing to its complexity. This study aimed to develop and assess an AI-driven diagnostic support system for magnetic resonance imaging (MRI)-based soft-tissue tumor diagnosis, potentially improving accuracy and aiding radiologists and orthopedic surgeons. Experienced orthopedic oncologists and radiologists annotated 720 images from 77 cases (41 benign and 36 malignant soft-tissue tumors). Eleven tumor subtypes were identified and classified into benign and malignant groups based on histological diagnosis. Utilizing the standard machine learning classifier pipeline, we examined and down-selected imaging protocols and their predominant radiomic features within the tumor's three-dimensional region to differentiate between benign and malignant tumors. Among the scan protocols, contrast-enhanced T1 weighted fat-suppressed images showed the most accurate classification based on radiomics features. We focused on the two-dimensional features from the largest tumor boundary surface and its neighboring slices, leveraging texture-based radiomic and deep convolutional neural network features from a pretrained VGG19 model. The test data comprised 44 contrast-enhanced images (22 benign and 22 malignant soft-tissue tumors) containing six malignant and five benign subtypes distinct from the training data. We compared expert and non-expert human performances against AI by assessing malignancy detection and the time required for classification. The AI model showed comparable accuracy (AUC 0.91) to that of radiologists (AUC 0.83) and orthopedic surgeons (AUC 0.73). Notably, the AI model processed data approximately 400 times faster than its human counterparts, showcasing its capacity to significantly boost diagnostic efficiency. We developed an AI-driven diagnostic support system for MRI-based soft-tissue tumor diagnosis. While additional refinement is necessary for clinical applications, our system has exhibited promising potential in differentiating between benign and malignant soft-tissue tumors based on MRI.

The Role of Affective Labour in Expertise: Bringing Emotions Back into Expert Practices

In recent years, a lot of scholarly attention has been devoted to how practices of digitalisation and datafication require medical professionals to work together with different stakeholders, and to how such collaborations shape expertise (Stevens, Wehrens, and de Bont 2020; Carboni et al. 2024). STS scholars have generally approached expertise as an epistemic and social endeavor, but they have tended to neglect the role affects and emotions play in its development and performance. In this paper, we provide a theoretical reflection on the relation between affective labour and expertise building upon Egher’s (2023) conceptualisation of expertise as a practical achievement realised through coordination and affective labour. Based on ethnographic fieldwork conducted in various medical settings, including digital pathology, psychiatry, and datafication in intensive care, we explore what types of affective labour are conducted in digital healthcare, by whom, and with what consequences. We show how affective labour mediates both epistemic and relational practices. We argue that different affects and emotions are mobilised in these practices, which impacts the development and effective performance of expertise.

Boardroom tenure, financial expertise and bank performance: a nonlinear dynamic framework

ABSTRACT This research investigates the influence of boardroom tenure and financial expertise on US bank performance in a nonlinear dynamic framework. To this aim, we utilize a sample of 305 banks during the 2010–2021 period, and we implement three empirical methodologies: GMM analysis, threshold technique, and quantile approach. Our findings reveal that the individual and interaction effects of boardroom financial expertise and tenure on bank performance are positive. Also, we provide evidence of an inverted U-shaped relation between boardroom tenure and bank performance, and between the board’s financial expertise and bank performance. Moreover, threshold analysis shows the presence of threshold effects in the relationship between boardroom tenure, expertise, and performance. Quantile results demonstrate that boardroom financial expertise improves the performance of high-performing banks. In contrast, boardroom tenure enhances the performance of low-performing banks but diminishes the performance of their high-performing counterparts.

Utilizing ChatGPT in Systematic Reviews and Meta-Analyses

The current uses, potential risks, and practical recommendations for using chat generative pre-trained transformers (ChatGPT) in systematic reviews (SRs) and meta-analyses (MAs) are reviewed in this article. The findings of prior research suggest that, for tasks such as literature screening and information extraction, ChatGPT can match or exceed the performance of human experts. However, for complex tasks such as risk of bias assessment, its performance remains significantly limited, underscoring the critical role of human expertise. The use of ChatGPT as an adjunct tool in SRs and MAs requires careful planning and the implementation of strict quality control and validation mechanisms to mitigate potential errors such as those arising from artificial intelligence (AI) 'hallucinations'. This paper also provides specific recommendations for optimizing human-AI collaboration in SRs and MAs. Assessing the specific context of each task and implementing the most appropriate strategies are critical when using ChatGPT in support of research goals. Furthermore, transparency regarding the use of ChatGPT in research reports is essential to maintaining research integrity. Close attention to ethical norms, including issues of privacy, bias, and fairness, is also imperative. Finally, from a human-centered perspective, this paper emphasizes the importance of researchers cultivating continuous self-iteration, prompt engineering skills, critical thinking, cross-disciplinary collaboration, and ethical awareness skills with the goals of: continuously optimizing human-AI collaboration models within reasonable and compliant norms, enhancing the complex-task performance of AI tools such as ChatGPT, and, ultimately, achieving greater efficiency through technological innovative while upholding scientific rigor.

Ultrasomics differentiation of malignant and benign focal liver lesions based on contrast-enhanced ultrasound

ObjectivesTo establish a nomogram for differentiating malignant and benign focal liver lesions (FLLs) using ultrasomics features derived from contrast-enhanced ultrasound (CEUS).Methods527 patients were retrospectively enrolled. On the training cohort, ultrasomics features were extracted from CEUS and b-mode ultrasound (BUS). Automatic feature selection and model development were performed using the Ultrasomics-Platform software, outputting the corresponding ultrasomics scores. A nomogram based on the ultrasomics scores from artery phase (AP), portal venous phase (PVP) and delayed phase (DP) of CEUS, and clinical factors were established. On the validation cohort, the diagnostic performance of the nomogram was assessed and compared with seniorexpert and resident radiologists.ResultsIn the training cohort, the AP, PVP and DP scores exhibited better differential performance than BUS score, with area under the curve (AUC) of 84.1-85.1% compared with the BUS (74.6%, P < 0.05). In the validation cohort, the AUC of combined nomogram and expert was significantly higher than that of the resident (91.4% vs. 89.5% vs. 79.3%, P < 0.05). The combined nomogram had a comparable sensitivity with the expert and resident (95.2% vs. 98.4% vs. 97.6%), while the expert had a higher specificity than the nomogram and the resident (80.6% vs. 72.2% vs. 61.1%, P = 0.205).ConclusionsA CEUS ultrasomics based nomogram had an expert level performance in FLL characterization.

Why the post-identification era is long overdue: Commentary on the current controversy over forensic feature comparison as applied to forensic firearms examination

In this commentary, we critically review recurring arguments for and against the discipline of forensic feature comparison as applied to firearms examination from various commentators within and outside forensic science. One of the mainstream criticisms that we address, among others, is that the field cannot demonstrate sufficient proficiency and robustness based on empirical (i.e., black-box) studies. While the lack of empirically demonstrated examiner proficiency is a valid concern and a powerful concept in the short term (e.g., in admissibility proceedings), many critics reduce their discussion of forensic feature comparison solely to the need to measure and demonstrate proficiency through error rates. However, the exclusive focus on aggregate expert performance metrics, here referred to as examiner diagnosticism, remains a surface-level perspective. It provides an incomplete account of the field because these metrics do not represent—but are often confused with—the notion of the evidentiary value of findings, i.e., observations made on examined items in individual cases. We argue that examiner diagnosticism should be contrasted and complemented with the notion of feature selectivity, i.e., the diagnostic capacity of observed marks and features on examined items. We argue that forensic scientists should report and be probed on their ability to quantify feature selectivity (i.e., the probative value of findings). By ceasing to express source attribution opinions (identification/individualisation), which are now widely exposed as unscientific, the forensic feature comparison disciplines could move further into the long-awaited post-identification era pioneered by other fields such as forensic genetics.

Horizontal Transition of Expertise

Expert performance in a domain is often defined as maximal adaptation to stable task constraints. This definition is useful when analyzing the vertical transition when novices become experts. However, many workplaces undergo considerable changes and, thus, task constraints change as well. In this paper, a complementary conceptualization of expertise is offered, one that focuses on expert performance as recurring adaptation to dynamic task constraints. This definition is useful when analyzing the horizontal transition when experts adapt to dynamically changing work contexts. Using the documentary method, the aim of the present study was to analyze cases of horizontal transitions based on qualitative biographical interview data from five experts reconstructing different types of adaptation to technological change in their domains that they have experienced. Implications for studying horizontal transitions at dynamic worksites are discussed.

Flow and Athletic Experts

AbstractPhilosophers frequently refer to ‘flow’ when discussing skill and expertise. This is especially clear when considering the ‘flow argument’ advocated by the Automatic view. Although the Automatic view is a minority position today, many philosophers and scientists endorse the flow argument or central parts of it. The present paper argues that the argument is implausible in the light of current psychological research. It proceeds in two steps. First, the paper argues that the psychological research does not suggest that flow is the only type of psychological state characteristic of optimal expert performance. Experts undergo different states, notably the state of clutch. Second, the paper argues that the psychological research does not suggest that flow states are devoid of conscious supervision in the form of conscious control, self-awareness, or task-related thoughts. Experts execute multiple conscious operations involved in attention control, motivational self-talk, and certain psychological mindsets while being in flow. These considerations oppose the two premises of the flow argument. Moreover, they correct misunderstandings of flow in the broader philosophical and scientific literature.

Experimental insights into cognition, motor skills, and artistic expertise in Paleolithic art

The production of Paleolithic art represents one of the most intricate technical and cognitive endeavors of Homo sapiens, marked by its profound antiquity and vast temporal and spatial framework. Despite its significance, there have been no prior studies aimed at understanding the cognitive and motor skills linked to the creation of realistic images characteristic of this artistic cycle. This research integrates archaeology and experimental psychology, premised on the assumption that the neurological basis of Anatomically Modern Humans has not changed substantially since the Upper Paleolithic. This work employs an innovative interdisciplinary approach, utilizing psychometric tests and drawing and engraving tasks monitored by motion-sensing gloves, to compare the performance of experts and non-experts in visual arts when faced with challenges akin to those of Upper Paleolithic artistic production. The results revealed that expertise in visual arts is linked to enhanced spatial abilities and specific patterns in drawing from memory. Additionally, both experts and non-experts displayed similar motor skills when engraving using Paleolithic techniques, suggesting that these techniques required specialized training in the contemporary experts. In conclusion, this research deepens our understanding of the processes involved in Upper Paleolithic artistic production.

Goal-directed learning is multidimensional and accompanied by diverse and widespread changes in neocortical signaling.

New tasks are often learned in stages with each stage reflecting a different learning challenge. Accordingly, each learning stage is likely mediated by distinct neuronal processes. And yet, most rodent studies of the neuronal correlates of goal-directed learning focus on individual outcome measures and individual brain regions. Here, we longitudinally studied mice from naïve to expert performance in a head-fixed, operant conditioning whisker discrimination task. In addition to tracking the primary behavioral outcome of stimulus discrimination, we tracked and compared an array of object-based and temporal-based behavioral measures. These behavioral analyses identify multiple, partially overlapping learning stages in this task, consistent with initial response implementation, early stimulus-response generalization, and late response inhibition. To begin to understand the neuronal foundations of these learning processes, we performed widefield Ca2+ imaging of dorsal neocortex throughout learning and correlated behavioral measures with neuronal activity. We found distinct and widespread correlations between neocortical activation patterns and various behavioral measures. For example, improvements in sensory discrimination correlated with target stimulus evoked activations of response-related cortices along with distractor stimulus evoked global cortical suppression. Our study reveals multidimensional learning for a simple goal-directed learning task and generates hypotheses for the neuronal modulations underlying these various learning processes.

Evaluations of solo piano performances: The role of performing with and without a musical score

The act of performing without a musical score is common and expected in many contexts. The purpose of this study was to examine the effects of performing with and without a musical score on listeners’ evaluations of solo pianists. We also investigated whether the effects of score use differed according to the perceived expertise of the performer. Participants ( N = 69) were collegiate music students who observed two video recordings of a pianist performing an excerpt from Schubert’s Impromptu Op. 90, No. 2. In one of the videos, the pianist used a musical score and in the other video the pianist did not use a score. One group of participants was informed at the outset that the pianist was an expert performer and the other group was told that the pianist was a student auditioning for a graduate program as a piano major. Participants rated each performance and indicated which recording they preferred. Results indicated no significant difference in ratings by performance condition overall, but significant interaction effects indicated that the effect of performance condition was dependent upon both the perceived expertise of the performer and presentation order. Chi-square results indicated no significant differences in participants’ preferences between conditions.

INVESTIGATING THE IMPACT OF AUDIT COMMITTEE ATTRIBUTES ON COMPANY PERFORMANCE: A MALAYSIAN FOCUS

Background and Purpose: This study aims to examine the relationship between selected audit committee attributes (gender diversity, size, meeting, independence, expertise) and company performance that are assessed as return on equity (ROE) and return on asset (ROA). As one of the primary corporate governance mechanisms facilitating corporate efficiency, audit committees positively impact company performance. Methodology: This quantitative cross-sectional study underpinned the agency theory through the target population: public listed companies under the Bursa Malaysia. Secondary data were gathered from the corporate annual report of 106 companies and Eikon database between 2018 and 2020 and subsequently analyzed with Pearson Correlation and Multiple Linear Regression using SPSS. Findings: Resultantly, a positive gender diversity-size correlation was identified with company performance as opposed to meeting, independence, and expertise. Furthermore, a significant gender diversity-size relationship was determined with company performance while no substantial counterparts between (i) meeting and independence and (ii) expertise and company performance were identified for the selected public listed companies. In this vein, audit committee attributes (gender diversity and size) could be utilized as a mechanism for high company performance. Contributions: The study outcomes offer useful insights into optimizing current corporate governance structures and regulations for relevant authorities. Keywords: Audit committee, corporate governance, company performance, Malaysia. Cite as: Gregory, V., Yusoff, H., Tapsir, R., &amp; Fauzi, H. (2024). Investigating the impact of audit committee attributes on company performance: a Malaysian focus. Journal of Nusantara Studies, 9(2), 185-217. http://dx.doi.org/10.24200/jonus.vol9iss2pp185-217