About

Aaron Mueller is an Assistant Professor of Computer Science at Boston University. His research centers on natural language processing and machine learning, with a focus on understanding, improving, and precisely controlling the capabilities and inner workings of systems that can learn and use human language.

Research topics

• Medicine
• Internal medicine
• Intensive care medicine
• Emergency medicine
• Virology
• Immunology
• Pediatrics

Selected publications

• Assessing The Predictive Value Of Lipoprotein(a) And C-reactive Protein For Detecting Subclinical Atherosclerosis

  Journal of cardiovascular computed tomography · 2025-07-01

  article1st authorCorresponding
  PublisherDOI

• Moving the Goalposts: Need to Adjust for Vasoactive Inotropic Support when Interpreting Hemodynamics Effects of the Impella 5.5 in Cardiogenic Shock Patients

  Journal of Cardiothoracic and Vascular Anesthesia · 2025-10-15

  letterOpen access
  PublisherOA PDFDOI

• Refining Cardiovascular Risk: Reducing Events Using Imaging-Based Treatment Strategies

  Journal of cardiovascular computed tomography · 2025-07-01

  article1st authorCorresponding
  PublisherDOI

• Longitudinal impact of cross-clamp duration on postoperative sleep disturbance and quality of life in elderly cardiac surgery patients: a secondary analysis of the MINDDS trial

  Frontiers in Anesthesiology · 2025-01-22

  articleOpen access

  Objectives This study aimed to assess the enduring impact of cross-clamp duration on postoperative sleep disturbance and functional outcomes (up to 180 days) in cardiac surgery patients. Design This is a secondary analysis of data from a randomized, double-blind trial comparing dexmedetomidine to placebo for delirium prevention (Minimizing ICU Neurological Dysfunction with Dexmedetomidine-induced Sleep). Setting Data from patients recruited at a tertiary medical center in Boston, Massachusetts, between March 2017 and February 2022 were analyzed in January 2024. Participants The study included 394 patients aged ≥60 who underwent cardiac surgery with cardiopulmonary bypass. Interventions The primary exposure was cross-clamp time, while secondary exposures included surgical type [isolated coronary artery bypass graft (CABG) or not] and dexmedetomidine randomization. Measurements and main results The primary outcome was sleep quality, assessed using the PROMIS Sleep Disturbance questionnaire at 30, 90, and 180 days postoperatively. Secondary outcomes encompassed cognitive function and health-related quality of life in various domains. Sleep quality, measured by PROMIS scores, showed improvement over time, and did not differ based on cross-clamp duration (MD 0.74 points, 95% CI: −0.57, 2.07), procedure type (MD 2.14 points, 95% CI: 0.29, 3.99), or dexmedetomidine (MD 0.9 points, 95% CI: −1.33, 1.5). However, isolated CABG patients reported sleep disturbance at all time points. Notably, extended cross-clamp time (>90 min) significantly worsened the trajectories of mental health (90-day: MD −2.37 points, 95% CI: −4.35, −0.39; 180-day: MD −2.68 points, 95% CI: −4.62, −0.73) and applied cognition (180-day: MD: −2.59 points, 95% CI: −4.49, −0.68). Conclusion Regardless of the duration of the cross-clamp, sleep quality tends to improve over time following cardiac surgery. However, cross-clamp times that last longer than 90 min have been identified as a risk factor for self-reported declines in mental health and applied cognition.

  PublisherDOI

• Association of Anesthesia Type and Vasoconstrictor Use with Intranodal Lymphangiography Duration: A Multicenter Retrospective Cohort Study

  Journal of Vascular and Interventional Radiology · 2025-08-26

  article
  PublisherDOI

• Anesthesia-induced electroencephalogram oscillations and perioperative outcomes in older adults undergoing cardiac surgery

  Journal of Clinical Anesthesia · 2025-02-07 · 5 citations

  articleOpen access
  PublisherOA PDFDOI

• Response to the Letter to the Editor: “Signs or symptoms of suspected preeclampsia – A retrospective National database study of prevalence, costs, and outcomes”

  Pregnancy Hypertension · 2025-01-10

  letter
  PublisherDOI

• Associations between Aβ40, Aβ42, and tau and postoperative delirium in older adults undergoing cardiac surgery

  Journal of Neurology · 2025-05-12 · 10 citations

  article
  PublisherDOI

• Long-chain polyunsaturated fatty acid lipid and oxylipin alterations in postoperative delirium after cardiac surgery

  Journal of Lipid Research · 2025-12-06

  articleOpen access

  Lipids play a crucial role in signaling, membrane dynamics, and inflammatory regulation, yet their involvement in postoperative delirium pathogenesis remains unclear. This study examined serum lipidomic alterations in postoperative delirium and assessed the effects of dexmedetomidine treatment on these changes. Lipidomic profiling was conducted at baseline and postoperative day 1 in two independent cohorts of cardiac surgery patients. Mass spectrometry-based shotgun lipidomics and targeted lipid analyses were used to assess lipidomes and oxylipins, respectively. Cardiac surgery was associated with decreased serum lysophospholipids. Postoperative delirium was associated with increased long-chain polyunsaturated fatty acid phospholipids, particularly phosphatidylethanolamines, and elevated oxylipins. Dexmedetomidine, a potential delirium-mitigating medication, reduced long-chain polyunsaturated fatty acid phospholipids. These findings highlight lipid modulation as a potential target for postoperative delirium prevention.

  PublisherDOI

• Detection of Venous Air Embolism in Nonsitting Craniotomy for Tumor Patients: A Retrospective Case Series

  Journal of Neurosurgical Anesthesiology · 2025-07-31 · 1 citations

  articleCorresponding

  BACKGROUND: Venous air embolism (VAE) occurs when air enters the venous circulation. During nonsitting craniotomies with elevated VAE risk due to proximity to a venous sinus, our institutional practice is to employ precordial Doppler ultrasound (PDU) and transesophageal echocardiography (TEE) for monitoring, as well as central venous catheterization (CVC) for aspiration. We utilized an electronic medical record (EMR) database to assess the frequency of VAE occurrence, its clinical detection, and the use of VAE-specific monitoring modalities. METHODS: EMR review identified all patients who underwent nonsitting craniotomies for an intracranial tumor. To identify episodes of VAE occurrence, the EMR was screened for intraoperative VAE events as determined by clinical diagnosis (cVAE) as well as an EtCO 2 drop >20% over a 2-minute interval, concerning for suspected VAE (sVAE). To identify patients who had VAE-specific monitoring, the EMR was scanned for placement of a CVC, TEE, or PDU. RESULTS: Three thousand nine hundred forty-five patients underwent a craniotomy for resection of tumor, and 3531 met study inclusion criteria. There were 14 episodes of intraoperative VAE diagnosed by a clinician (cVAE) and 86 episodes of suspected VAE (sVAE) based on review of anesthesia records for significant changes in EtCO 2 . There were 261 cases that used VAE-specific monitoring, with minimal overlap with sVAE cases. CONCLUSIONS: We identified 100 episodes of VAE, diagnosed either clinically (cVAE) or by abrupt EtCO 2 decrease (sVAE). Our data suggest that VAE in nonsitting craniotomy often occurs in instances where VAE-specific monitoring modalities are not used, and that our ability to preoperatively identify neurosurgical cases where VAE may occur is limited.

  PublisherDOI

Frequent coauthors

• Sarosh Rana

  University of Chicago Medical Center

  235 shared

• Daniel Talmor

  Beth Israel Deaconess Medical Center

  173 shared

• Avery Tung

  University of Chicago

  143 shared

• Shahzad Shaefi

  Beth Israel Deaconess Medical Center

  121 shared

• Sajid Shahul

  University of Chicago

  120 shared

• Joana Lopes Perdigao

  95 shared

• Brian O’Gara

  Harvard University

  86 shared

• Valerie Banner‐Goodspeed

  Beth Israel Deaconess Medical Center

  81 shared