About

Yakun Zhang is an Assistant Professor in the Department of Soil and Environmental Sciences at the University of Wisconsin-Madison, affiliated with the College of Agricultural & Life Sciences. His research integrates soil and agricultural sciences with advanced artificial intelligence techniques to support sustainable agricultural and natural resources management. His work includes spectral analysis for rapid soil health assessment and carbon measurement, dynamic soil survey under natural and anthropogenic disturbances such as wildfires and urbanization, and the application of big data and geospatial machine learning. Zhang's research also focuses on understanding soil-landscape relationships and their evolution under climate change and human activities, as well as polar soil and earth systems modeling.

Research topics

• Environmental science
• Soil science
• Computer Science
• Geology
• Remote sensing
• Chemistry
• Mathematics
• Materials science
• Geotechnical engineering
• Biology
• Environmental planning
• Environmental resource management
• Agroforestry
• Business
• Composite material
• Statistics
• Mineralogy
• Ecology
• Agronomy
• Geomorphology
• Chromatography
• Optics
• Physics
• Geography

Selected publications

• Study on the corrosion damage behavior and anti-corrosion methods of mining anchor cables in high-mineralized water environments

  Construction and Building Materials · 2026-05-04

  articleOpen access

  Corrosion failure of anchor cables caused by highly mineralized water in deep coal mine roadways has become increasingly severe, often leading to local roof block falls or even large-scale rock collapses, thereby threatening safe mining operations. To address this issue, laboratory salt-spray accelerated corrosion tests were conducted on anchor cables. Corrosion damage behavior under a 5% chloride ion concentration was systematically investigated using metallographic analysis, scanning electron microscopy (SEM), and energy-dispersive spectroscopy (EDS), and the associated damage mechanisms were elucidated. Results show that in a high-mineralized water environment, Cl⁻ readily induces uniform corrosion and pitting corrosion of anchor cables. At early corrosion stages, a yellow-brown rust layer forms on the cable surface and gradually develops into a loose, porous structure as exposure time increases, while corrosion rate exhibits a nonlinear trend, decreasing first and then increasing. As Cl⁻ penetrates into the matrix, Fe atoms around inclusions preferentially dissolve due to lattice energy differences, promoting pit initiation. Metallographic and SEM-EDS analyses indicate that corrosion products mainly consist of oxides and sulfide-type inclusions, and corrosion progression causes grain boundary degradation, pore formation, and preferential corrosion occurs at the inclusion-matrix interface, which significantly weaken anchor cable strength and ductility. Through comparative analysis of corrosion morphology and post-corrosion physical and mechanical properties of anchor cables with different coatings, aluminum-zinc coating is identified as an effective anti-corrosion process. In addition, effects of post-corrosion cable morphology on anchorage load-bearing capacity, and further clarifies the degradation laws of anchor cables under single-factor and coupled effects, as well as their engineering applicability are examined. These findings provide a theoretical basis and technical support for corrosion protection of anchor cables in coal mines. • Elucidated corrosion mechanisms and microstructural characteristics of anchor cables in severe corrosive conditions. • Evaluated and optimized four anti-corrosion processes and their corresponding control methods. • Revealed nonlinear characteristics of anchor-cable corrosion rate and clarified influence of elemental distributions. • Summarized effects of stress, corrosion, and their stress–corrosion coupling on anchor-cable corrosion.

  PublisherDOI

• Quantifying source-specific human and ecological risks of heavy metals in riparian wetlands using integrated chemical speciation and APCS-MLR modeling

  Environmental Pollution · 2026-05-18

  article
  PublisherDOI

• 0069 Sleep Efficiency and Memory-Related Brain Functional Connectivity

  SLEEP · 2026-05-01

  article1st authorCorresponding

  Abstract Introduction Sleep plays a critical role in memory consolidation, yet the impact of chronic sleep deficiency on functional connectivity (FC) in key memory-related regions, such as the hippocampus (HPC) and posterior cingulate cortex (PCC), remains unclear. Traditional BOLD fMRI suffers from signal dropout near air-tissue interfaces, limiting its ability to reliably assess FC in the HPC. Dynamic arterial spin labeling (DASL) offers a potential solution by providing more robust perfusion-based measures of FC. In this preliminary study, we examine how sleep efficiency relates to HPC and PCC FC using both BOLD and DASL imaging. Methods Fourteen patients (60±16 years; 8 females) who underwent clinical polysomnography (PSG) were recruited, and sleep efficiency was scored by certified technicians. MRI data were acquired on a 3T Siemens Prisma and included a 5-minute BOLD scan, a 5-minute DASL scan using background-suppressed pseudo-continuous arterial spin labeling (PCASL), and a high-resolution T1-weighted structural image. Seed-based FC for the PCC and left HPC (LHPC) was calculated using voxel-wise Pearson correlations followed by Fisher z-transformation. Associations between sleep efficiency and FC were assessed using multiple linear regression with ANCOVA controlling for global signals, along with correction applied for multiple comparisons. Results Sleep efficiency was significantly associated with FC in both the PCC and LHPC. Using BOLD, higher sleep efficiency was associated with increased PCC FC in the anterior cingulate and medial orbitofrontal cortex, regions within the default mode network (DMN). Using DASL, higher sleep efficiency was associated with increased PCC FC in the left insula, a key node of the salience network (SN), as well as increased LHPC FC in the right supramarginal and superior temporal region, which are components of the DMN. Conclusion Higher sleep efficiency was associated stronger FC between memory-related regions and major large-scale networks, including the DMN and SN. These findings suggest that reduced sleep efficiency may disrupt network-level connectivity patterns supporting memory and attention. Such disruptions may contribute to the neural mechanisms linking poor sleep to impaired memory consolidation and increased vulnerability to memory decline and dementia. Future work will integrate automating sleep-efficiency estimation into our modeling framework. Support (if any) Transdisciplinary Areas of Excellence seed grant at Binghamton University

  PublisherDOI

• Sophoridine derivatives: Synthesis, anti–Spodoptera litura activity, and plant growth–promoting effects

  Process Biochemistry · 2026-02-25

  article
  PublisherDOI

• 0363 Improving Automated Sleep Staging with Deep Learning: Learnable PSG Channel Encoding

  SLEEP · 2026-05-01

  article

  Abstract Introduction Sleep scoring is essential for clinical sleep diagnostics, and recent advances in deep learning have accelerated and standardized this process. Transformer-based models such as SleepTransformer have achieved state-of-the-art performance, and our previous work, FlexSleepTransformer, further improved scoring accuracy and cross-dataset generalizability by fusing information from multiple PSG channels. However, these methods rely on fixed channel encodings that fail to capture the spatial organization of PSG electrodes, limiting their ability to model inter-channel relationships effectively. To address this limitation, we introduced a learnable 2D PSG channel encoding that explicitly represents spatial structure and integrated it into FlexSleepTransformer, leading to improved performance across multiple datasets. Methods A total of 543 subjects from seven independently acquired datasets were included. For each dataset, subject-level 5-fold cross-validation was performed to prevent data leakage. The baseline model followed the two-level sequence-to-sequence SleepTransformer architecture, which processed intra-epoch information and inter-epoch temporal context in a manner consistent with human scoring guidelines. Because the datasets differed in their PSG channel configurations, the model required a way to recognize each channel’s spatial origin. Traditional fixed channel encodings identified separate channels but failed to capture spatial relationships. To overcome this limitation, we introduced a learnable 2D PSG positional encoding that allowed the model to autonomously learn spatially informed representations of signals from different brain regions. Three models were evaluated across all datasets: (1) no channel encoding, (2) fixed channel encoding, and (3) the proposed learnable 2D channel encoding. Results Across all seven datasets, the proposed learnable 2D channel encoding achieved the highest average accuracy (82.84%±2.08%), outperforming both the no-encoding model (81.6%±1.79%) and the fixed-encoding model (82.35%±1.53%). Statistical comparisons further showed that the learnable encoding significantly outperformed the no-encoding baseline on six datasets and outperformed the fixed encoding on four datasets, demonstrating its consistent advantage across diverse data sources. Conclusion We introduced a learnable 2D EEG channel encoding for Transformer-based sleep staging and successfully incorporated it into FlexSleepTransformer. Results across seven datasets demonstrated consistent and significant performance improvements, highlighting the strong potential of this approach for deployment in real clinical workflows. Support (if any) None

  PublisherDOI

• Distinguishing Soil Horizons of Alfisols Using <scp>MIR</scp> , Vis– <scp>NIR</scp> , and <scp>XRF</scp> Spectra

  European Journal of Soil Science · 2026-04-28

  articleOpen accessSenior author

  ABSTRACT We studied horizonation of 150 Alfisols on a research farm in Wisconsin, USA. Most of these Alfisols had developed in loess covering dense glacial till; common horizons were an Ap over a glossic horizon (E/Bt or Bt/E) over a Bt covering the glacial till (2Bt). In addition to field identification, we distinguished the horizons by combining soil properties and spectra (Vis–NIR, MIR, and XRF) using a random forest (RF) model. A total of 503 soil samples from 150 pedons were analyzed. Soil samples were scanned using visible–near infrared (Vis–NIR) (350–2500 nm), mid‐infrared (MIR) (4000–600 cm −1 ), and X‐ray fluorescence (XRF) (0–50 keV) spectrometers. Higher silt, pH, soil organic carbon (SOC), Ca, K, P, Si, Ti, Zn, and Zr contents were measured in the Ap horizon, whereas the glossic horizons were characterized by higher CIE L * (lightness) and b * (yellow‐blue) values. The underlying Bt horizon had higher clay, Al, and Fe contents, while the 2B(t) horizon was distinguished by higher sand, coarse fragments, and CIE a * values. The combination of soil properties and spectra (Vis–NIR, MIR, XRF) achieved a higher overall accuracy for predicting soil horizons compared to their individual use. The Ap horizons were best predicted using MIR or XRF (R 2 = 0.98), the glossic horizons using combined Vis–NIR and XRF spectra (R 2 = 0.98), the Bt horizons using MIR (R 2 = 0.90), whereas the 2B(t) horizons were best predicted using combined spectra (Vis–NIR + MIR + XRF) (R 2 = 0.50). In the absence of soil data, these spectra can be used with high accuracy to distinguish A and various B horizons of Alfisols.

  PublisherDOI

• Antibacterial and Antibiofilm Activity of <i>Clerodendron Cyrtophyllum</i> Turcz. Ethanolic Extracts against <i>Staphylococcus aureus</i> from Bovine Mastitis

  Journal of Microbiology and Biotechnology · 2025-11-27

  articleOpen access

  Guoshun Pei, Rong Cao, Qiguo Li, Yue Hu,Yakun Zhang, Yao Feng, Yanrong Zeng, Yudie Guo, Yehui Luo, Lina Liu , and Chengjian Tan. J. Microbiol. Biotechnol. 2025;35:1-11. https://doi.org/10.4014/jmb.2506.06007

  PublisherOA PDFDOI

• Abnormal Low Frequency Fluctuations and Functional Connectivity in Overactive Bladder Syndrome

  Proceedings on CD-ROM - International Society for Magnetic Resonance in Medicine. Scientific Meeting and Exhibition/Proceedings of the International Society for Magnetic Resonance in Medicine, Scientific Meeting and Exhibition · 2025-09-16

  article

  Motivation: Overactive Bladder (OAB) significantly affects patients' quality of life. However, the exact mechanism of brain-bladder control in urinary continence remains unclear. Goal(s): To investigate the abnormality of amplitude of low-frequency fluctuations (ALFF) and functional connectivity (FC) during the bladder filling in OAB patients. Approach: Blood oxygenation level-dependent (BOLD) images were acquired at bladder filling volumes of 0, 50, 100, 200, 350 and 500mL. Results: We observed comprised ALFF in the orbitofrontal cortex and medial prefrontal cortex (mPFC), along with disrupted FC between posterior cingulate cortex (PCC) and mPFC and between mPFC and pDMN in OAB patients during bladder filling. Impact: We provide new insights into the roles of FC and ALFF in controlling urinary incontinence. This study demonstrated that FC and ALFF may serve as potential biomarkers for monitoring treatment effects in OAB patients.

  PublisherDOI

• Age-associated alterations of brain networks using dynamic ASL

  Proceedings on CD-ROM - International Society for Magnetic Resonance in Medicine. Scientific Meeting and Exhibition/Proceedings of the International Society for Magnetic Resonance in Medicine, Scientific Meeting and Exhibition · 2025-09-16

  article1st authorCorresponding

  Motivation: Motion may affect age-related discrepancy in resting-state functional connectivity (rsFC) in BOLD imaging studies. Dynamic arterial spin labeling (DASL) offers an alternative, detecting large-scale network rsFC with fewer motion artifacts. However, the aging effects on DASL rsFC remain unexplored. Goal(s): We examined how brain networks evolve with age using DASL. Approach: We used a data-driven approach to measure rsFC within and between brain networks in 131 adults. Results: Older age correlated with reduced intra-network rsFC in the ventral stream (VSN), sensorimotor, left frontoparietal (LFPN), right frontoparietal, default mode (DMN), salience (SN), and frontal networks, and decreased inter-network rsFC between LFPN-SN, VSN-SN, and DMN-SN. Impact: The rsFC of DASL may offer additional knowledge about the aging brain in the healthy adults and increase our insight into the neural basis of neurodegenerative disease.

  PublisherDOI

• ALK-Targeted Therapy: Resistance Mechanisms and Emerging Precision Strategies

  Current Issues in Molecular Biology · 2025-11-27 · 3 citations

  articleOpen access1st author

  Anaplastic lymphoma kinase (ALK), a member of the receptor tyrosine kinase family, plays a central oncogenic role in the initiation and progression of diverse malignancies. Aberrant ALK activation generally results from structural alterations or dysregulated expression, leading to persistent activation of downstream signaling pathways that drive tumor cell proliferation, survival, and metastasis. ALK gene abnormalities predominantly encompass fusions, point mutations, and amplifications, with EML4-ALK-positive non–small cell lung cancer representing a canonical example. The advent of ALK-targeted inhibitors has constituted a major therapeutic milestone for ALK-positive tumors. From first-generation Crizotinib to third-generation Lorlatinib, successive agents have been refined for target selectivity, central nervous system penetration, and coverage of resistance-associated mutations, substantially improving patient survival and intracranial disease control. Nonetheless, the emergence of acquired resistance remains an overarching challenge, mediated by secondary kinase domain mutations, activation of bypass signaling pathways, and tumor phenotypic transformation. This review presents an integrative synthesis of ALK-targeted therapeutic developments, elucidates underlying resistance mechanisms, and surveys emerging strategies, providing a comprehensive perspective on current advances and future directions in precision management of ALK-driven malignancies.

  PublisherOA PDFDOI

Frequent coauthors

• Ehsan Samei

  Duke University Hospital

  52 shared

• Alfred E. Hartemink

  48 shared

• W. Paul Segars

  Duke University

  34 shared

• Guoping Zhao

  20 shared

• Jingyi Huang

  University of Wisconsin–Madison

  17 shared

• Zailin Yang

  Chongqing University

  16 shared

• Gafur GÖZÜKARA

  University of Wisconsin–Madison

  15 shared

• Asim Biswas

  13 shared

Labs

• Soil and Environmental Sciences Analysis LabPI

  Not provided