Browse Articles
Discover research articles across all indexed journals
Formation of Ultra‐Narrowband SWIR J‐Aggregate Materials and Their Applications in Multispectral Optoacoustic Tomography Imaging
ABSTRACT Ultra‐narrowband materials provide exceptional spectral purity for high‐resolution imaging, precision sensing, and advanced photonic applications, yet molecular systems achieving ultra‐narrowband absorption in the short‐wave infrared (SWIR) remain exceedingly rare. This scarcity arises from the intrinsic trade‐off between long‐wavelength absorption, vibrational broadening, and the structural disorder that typically emerges during disordered aggregation. Here, we introduce SWIR cyanine dyes rationally engineered through controlled steric hindrance, sulfonate‐mediated electrostatic interactions, and π‐conjugation design to promote highly ordered J‐aggregation. The optimized dye, IRJ1089, assembles into remarkably stable and encapsulation‐free J‐aggregates, exhibiting high optoacoustic signal generation efficiency and ultra‐narrowband absorption at 1089 nm with a FWHM of 164 cm −1 (19.4 nm). Based on a structural analogue's (IRJ1021) crystallographic data and computational analyses, we propose that electrostatic interlocking between sulfonate groups and the polarized π‐backbone, together with favorable slip‐stacking geometry, promotes aggregate order and enhances the J‐aggregates’ stability. IRJ1089 J‐aggregates support high‐contrast multispectral optoacoustic tomography (MSOT) imaging in vasculature deep inside rats (∼1.2 cm), enable real‐time monitoring of antihypertensive drug‐induced aortic dilation, and allow crosstalk‐free SWIR multiplexed imaging for diagnosing colorectal cancer‐associated intestinal obstruction. These findings establish a generalizable blueprint for designing ultra‐narrowband J‐aggregating dyes and provide valuable insights into other SWIR photonic materials in multiple areas.
Erratum: VIKTORIA-1 Trial of Gedatolisib Plus Fulvestrant With or Without Palbociclib in Hormone Receptor–Positive/HER2−/PIK3CA Wild-Type Advanced Breast Cancer
Rethinking Optimal Therapy for Operable Pancreatic Cancer in the CASSANDRA Era
Influence of root inducers and substrates on the vegetative propagation of pitahaya species (Hylocereus spp.) under nursery conditions
Precursor Coordination Engineering Enables Epitaxial‐Level Carrier Densities in HgTe Colloidal Quantum Dots
ABSTRACT Surface trap‐induced carrier doping currently hinders colloidal quantum dots (CQDs) from approaching the electronic purity of epitaxial semiconductors, particularly in the mid‐wave infrared (MWIR) regime, where low dark noise is critical. Here, we demonstrate that the coordination geometry of the metal halide precursor strongly influences passivation efficacy, suppressing background carrier densities to an intrinsic level (∼10 14 cm −3 ). Mechanistic investigation reveals that the octahedral geometry of HgBr 2 facilitates the formation of a stable HgBr 2 (Olam) 4 complex, as traced by the correlated evolution of chemical shifts in 1 H NMR and diffraction patterns in powder X‐ray diffraction (pXRD). This behavior contrasts with the linear lattice of HgCl 2 , where the absence of comparable spectral evolution indicates much weaker ligand coordination. Consequently, the stable complex is structurally inherited by the HgTe CQDs, yielding a high surface halide coverage of ∼9 at.%, whereas HgCl 2 results in halide‐deficient surfaces (<1 at.%). This robust passivation effectively suppresses non‐radiative recombination, consistent with improved carrier decay dynamics in time‐resolved photoluminescence (TRPL). Cross‐verification via field‐effect transistor (FET), capacitance‐voltage (C‐V), and Hall measurements confirms substantially reduced background doping, highlighting coordination engineering as an effective strategy for achieving electronic purity in solution‐processed optoelectronics.
Li <sup>+</sup> ‐Mediated Topological Regulation of Aluminosilicate Glass Ceramics: Near‐Full Crystallinity for Multifunctional Optoelectronic Applications
ABSTRACT Developing glass ceramics (GCs) with ultrahigh crystallinity and excellent optical transparency remains a formidable challenge due to the intrinsic trade‐off between crystalline volume fraction and grain boundary light scattering. Herein, a Li + ‐ doping mediated topological network regulation strategy is developed to address this dilemma in MgO‐Al 2 O 3 ‐SiO 2 transparent GCs (TGCs). Combined experimental characterizations and molecular dynamics simulations confirm that Li + acts as a network modifier to relax the rigid tetrahedral framework, tailor Al coordination, and optimize crystallization kinetics. This boosts crystallinity from 5.6 to 97.0 vol% (near‐full crystallinity) while retaining high transparency via refractive index matching between the crystal and residual glass. Eu 2+ ‐activated high‐crystallinity TGCs exhibit superior multifunctional optical performance, including a high internal quantum efficiency of 63%, good resistance to thermal quenching (82% intensity retention at 150°C), and a high x‐ray light yield of 5740 photons/MeV. Their practical applicability is further validated for high‐power indoor/horticultural lighting and high‐resolution x‐ray scintillation. This work establishes a universal topological engineering paradigm for the rational design of glass network topologies, which provides a new solution to the transparency‐crystallinity trade‐off in GCs and paves the way for next‐generation high‐performance TGCs‐based optoelectronic devices.
Simulating expert oncologist-to-oncologist thinking and behavior: Multi-agent AI for precision oncology collaboration.
20 Background: Collaborative oncologist decision-making—via tumor boards, subspecialty consultation, and second opinions—frequently alters treatment recommendations, often in 20–40% of cases. However, maintaining expert synchrony is increasingly challenging due to rapid guideline updates, expanding therapeutic options, and workforce constraints, particularly in resource-limited settings. Single large language models (LLMs) demonstrate substantial rates of guideline-discordant recommendations in oncology. We developed a multi-agent AI system in which agents act on behalf of organ- and modality-specific oncologists to simulate expert cognitive reasoning and behavioral collaboration. Methods: The architecture comprises specialized agents representing medical, surgical, and radiation oncology operating through synchronized perceive–reflect–plan loops. A live knowledge layer integrates continuously updated clinical guidelines and regulatory approvals. A structured simulation engine governs deterministic inter-agent deliberation. Reinforcement learning using behavior cloning and multi-agent proximal policy optimization (MAPPO) promotes stable collaborative dynamics. The system was evaluated on 102 multi-institutional real-world breast cancer cases and 60 structured MEREDITH/AMIE-style benchmark tasks (total N = 162). Primary endpoints included guideline concordance, safety (major discordance), and comparative performance versus single-model baselines. Results: The multi-agent system achieved 95.1% guideline concordance (95% CI, 91.2%–98.4%) on structured benchmarks and 93.1% (95% CI, 89.1%–96.5%) on real-world cases. No major guideline-discordant recommendations were observed. Compared with single-model baselines, absolute improvements ranged from 15–47 percentage points (all P < .001). Inter-agent verification eliminated expert-rejected outputs observed in comparator systems. Conclusions: A multi-agent AI framework acting on behalf of organ-specific oncologists to simulate expert thinking and behavioral synchrony significantly improves guideline adherence while eliminating major discordant outputs. This architecture provides a scalable approach for precision oncology decision support and structured training in rapidly evolving clinical environments.
Comparative efficacy and safety of first-line maintenance therapies in advanced ovarian cancer: A biomarker-stratified Bayesian network meta-analysis of 16 phase 3 trials.
213 Background: The integration of poly(ADP-ribose) polymerase inhibitors (PARPi), immune checkpoint inhibitors (ICI), and anti-angiogenics has transformed first-line maintenance for advanced ovarian cancer (AOC). However, optimal regimens across homologous recombination deficiency (HRD) and BRCA mutation (BRCAm) profiles remain undefined. This network meta-analysis (NMA) evaluates the comparative efficacy and safety of emerging combinations against established standards. Methods: A systematic literature search identified Phase 3 randomized controlled trials evaluating first-line maintenance in AOC. Progression-free survival (PFS) log-hazard ratios and grade 3 adverse event (AE) rates were extracted. Bayesian NMAs were conducted, stratified by BRCAm, HRD-positive (HRd), and HRD-negative/proficient (HRp) status. Treatments were ranked using Surface Under the Cumulative Ranking (SUCRA) probabilities. Results: Sixteen phase 3 trials involving 11,482 patients were included. In the BRCAm network, PARPi monotherapies dominated efficacy, with olaparib (SUCRA 92%) and senaparib (SUCRA 88%) demonstrating superiority; adding ICI therapy (rucaparib+nivolumab) showed no synergistic benefit and ranked lowest (SUCRA 12%). In the HRd network, olaparib+bevacizumab achieved the highest probability of maximizing PFS (SUCRA 94%), outperforming triplet ICI combinations (durvalumab+olaparib+bevacizumab; SUCRA 78%). In the HRp network, individualized-dose niraparib (SUCRA 85%) and triplet therapies (SUCRA 74%) provided the most significant PFS advantages over baseline bevacizumab. Safety analyses revealed individualized PARPi dosing based on baseline body weight and platelets drastically improved tolerability rankings (SUCRA 62%) compared to fixed-dose administration (SUCRA 18%), mitigating severe hematological AEs. Conclusions: PARPi monotherapy remains the optimal frontline maintenance strategy for BRCAm AOC, as ICI combinations offer added toxicity without commensurate survival benefits. For HRd patients, combining a PARPi with an anti-angiogenic yields the highest efficacy. Across all subgroups, individualized PARPi dosing is critical for optimizing the therapeutic index.
Pd-Catalyzed Arylative Lossen Rearrangement: Synthesis of Secondary Amines from Aryl/Alkyl Carboxylic Acids and Aryl Halides
Electron‐Localized RuNi Nanosheet Assemblies Enable Low‐Overpotential Mg‐CO <sub>2</sub> Battery
ABSTRACT Mg‐CO 2 battery is emerging as a promising energy storage system that simultaneously converts CO 2 into value‐added products. However, its practical application is hindered by formation thermodynamically stable and electrically insulating discharge product MgCO 3 , which severely limits energy efficiency of Mg‐CO 2 battery. Herein, we report an electron‐localized Ru 61 Ni 39 nanosheet assembly catalyst that overcomes these limitations by precisely engineering the surface electronic structure for achieving precise tuning of product from MgCO 3 to MgC 2 O 4 of Mg‐CO 2 battery. We demonstrate that electron transfer from Ni to Ru creates the localized electron at the Ru sites, weakening MgC 2 O 4 binding and suppressing its conversion to MgCO 3 , thereby enabling reversible formation and decomposition of MgC 2 O 4 and mitigating cathode passivation. The Mg‐CO 2 battery incorporating the electron‐localized Ru 61 Ni 39 nanosheet assembly catalyst achieves an ultralow charge overpotential of 0.07 V and a record‐high energy conversion efficiency of 94.1%, with the stable cycling for over 580 h. In situ electrochemical spectroscopy and theoretical studies reveal that the electron‐localized between Ru and Ni stabilizes the product intermediates (C 2 O 4 2− ) and prevents the MgC 2 O 4 conversion to MgCO 3 .
Brain-structural differences underlying dialect competence in the bilingual network
Abstract Research in multilingualism has provided evidence for brain structural differences between monolinguals and bilinguals. Less is known about speakers who, apart from a standard language, also show competence in a variety (dialect) of this language (“bidialectals”). In these populations, cultural differences are minimized and language competence can be compared against a common ground. We hypothesize that bidialectals show brain-structural differences to non-bidialectals, and capitalize on one of the world’s largest dialect corpus available for German. A competent dialect group (N = 26) and a group without dialect competence (N = 23) were compared by using brain structural measures, including gray matter volume (GMV) and cortical thickness (CT). Results demonstrate a whole-brain group difference, seen for CT in right orbitofrontal cortex, and for GMV in bilateral middle temporal gyrus and bilateral insula. Notably, CT in right fusiform cortex as well as GMV in bilateral middle temporal gyrus and right insula co-varied with dialect competence differently for the two groups. The structural differences and covariations with dialect competence are discussed on the background of code switching and language control. The findings suggest that dialect competence may shape brain structure in ways similar to bilingualism.
Carbon Materials as Lithium‐Free Anode for Hybrid Lithium‐Ion/Metal Batteries: Mechanism, Design Strategy, In Situ Characterization, and Prospects
ABSTRACT Carbon based anodes that rely solely on Li‐ion storage are inherently limited in energy density, whereas systems dominated by Li‐metal storage suffer from poor reversibility and short cycle life. Hybrid Li‐ion/Li‐metal batteries (LIB/LMBs), enabled by carbon‐based Li‐free anodes, offer a promising pathway by integrating Li‐ion intercalation and Li‐metal plating/stripping within a single framework. Despite rapid progress, the practical implementation of hybrid LIB/LMBs remains hindered by low Coulombic efficiency, unstable solid electrolyte interphase, dendrite growth, and large volume fluctuations. These challenges are fundamentally associated with irreversible Li loss and interfacial instability, which limit long‐term cycling and energy efficiency. In this review, we systematically summarize recent advances in carbon‐based Li‐free anodes for hybrid LIB/LMBs, with a focus on (i) the hybrid storage mechanism, (ii) rational design of carbon materials, (iii) interface engineering strategies, (iv) mechanistic insights from in situ characterization, and (v) critical perspectives toward practical deployment. Particular emphasis is placed on the relationship between Coulombic efficiency, Li inventory retention, and system‐level performance. This review provides a unified framework for understanding hybrid storage chemistry and offers strategic guidance for the development of next‐generation high‐energy‐density batteries.
Tough and Healable Shape‐Memory Elastomers Enabled by Dynamic Hyperbranched Topological Networks
ABSTRACT Integrating rapid self‐healability and shape memory into tough polyurethane elastomers is highly desirable for applications of artificial muscles and soft actuators. However, this combination is fundamentally limited by an inherent structural trade‐off between strong phase separation and dynamic interactions. Herein, we develop a dynamic hyperbranched topology strategy to engineer gold nanoparticle‐crosslinked polyurethane elastomers featuring a hierarchical structure that includes dynamic topological networks, enhanced hard domains with strong hydrogen bonds, and nano‐assemblies phase‐separated from an amorphous matrix in which soft segments are prepacked for ultrasensitive strain‐induced crystallization (SIC). Reversible dissociation/combination of dynamic bonds and progressive rupture of hard domains and assemblies contribute to energy dissipation, while ultrasensitive SIC enables network self‐reinforcement, affording a high strength of 53 MPa, a toughness of 483.3 MJ m −3 , and a fracture energy of 352.5 kJ m −2 . Meanwhile, abundant dynamic bonds enable rapid self‐healability of the elastomer with a healing efficiency of 93.5% upon 2‐min near‑infrared irradiation. Accompanied by sensitive SIC‐induced network fixation, these elastomers show remarkable shape memory with a shape fixation ratio of 91% and a shape recovery ratio of 94.7%. Based on the unprecedented combinational properties, this work establishes a design paradigm for hierarchically structured elastomers that reconcile high mechanical performance with smart stimulus responsiveness.
Mortality trends and disparities in bone-metastatic prostate cancer in the United States, 1999–2024: A CDC WONDER analysis.
189 Background: In United States, mortality trends for bone-metastatic prostate cancer remain incompletely characterized. This study examines longitudinal trends and demographic and geographic disparities using a multiple cause-of-death approach. Methods: Mortality data for adults ≥25 years were obtained from CDC WONDER (1999–2024). Deaths required prostate cancer (ICD-10: C61) as underlying cause and bone metastasis (C79.5) as a contributing cause. Age-adjusted mortality rates (AAMRs) per 100,000 were stratified by year, age, race/ethnicity, census region, and urbanization. Joinpoint regression estimated APC and AAPC with 95% CIs; p < 0.05 was considered significant. Results: A total of 73,429 deaths were identified (Table 1). Overall AAMR rose from 1.10 (95% CI: 1.05–1.15) in 1999 to 2.12 (2.06–2.17) in 2024 (AAPC: 2.61; 95% CI: 1.66–3.56; p < 0.001). Four trend segments were identified: decline from 1999–2006 (APC: −5.57), modest rise from 2006–2013 (APC: 2.92), sharp increase from 2013–2018 (APC: 12.91), and continued rise from 2018–2024 (APC: 4.02). Adults aged 65–85+ had the highest burden (AAMR 5.09 to 9.79; AAPC: 2.73; p < 0.001). Non-Hispanic (NH) Black individuals had the highest 2024 AAMR (3.09), compared to NH White (2.12) and Hispanic/Latino (1.62); NH Black AAPC was not statistically significant (1.11; p = 0.37), while NH White showed significant increase (AAPC: 3.16; p < 0.001). The West had the largest absolute increase (0.85 to 2.53; AAPC: 4.75; p < 0.001). Non-metropolitan areas had higher 2024 AAMRs than metropolitan areas (1.94 vs. 1.73). Most deaths occurred at home (48.51%); only 10.47% occurred in hospice facilities. Conclusions: Bone-metastatic prostate cancer mortality increased over 25 years, with persistent disparities by age, race/ethnicity, region, and urbanization. These findings may inform targeted public health planning and resource allocation. Bone-metastatic prostate cancer mortality rates (per 100,000) by demographic groups, 1999 and 2024. Variable Deaths AAMR (95% CI), 1999 AAMR (95% CI), 2024 Overall 73,429 1.10 (1.05–1.15) 2.12 (2.06–2.17) NH Black 10,417 2.47 (2.21–2.73) 3.09 (2.89–3.31) NH White 56,125 1.00 (0.95–1.05) 2.12 (2.06–2.18) Hispanic/Latino 5,088 1.23 (0.98–1.48) 1.62 (1.47–1.78) Midwest 16,594 1.16 (1.06–1.26) 2.29 (2.17–2.42) Northeast 9,092 1.04 (0.94–1.14) 1.27 (1.17–1.37) Metropolitan 41,470 1.07 (1.02–1.13) 1.73 (1.68–1.79)* Non-metropolitan 9,987 1.19 (1.08–1.31) 1.94 (1.81–2.07)* NH = Non-Hispanic. AAMRs per 100,000 population, age-adjusted to the 2000 U.S. standard population. *Urbanization data available through 2020 only.
INAVO120 phase 3 trial: Subgroup analyses of Asian patients (pts) with <i>PIK3CA</i> -mutated (mut), hormone receptor-positive, HER2-negative (HER2–), endocrine-resistant advanced breast cancer (aBC) treated with inavolisib (INAVO)/placebo (PBO) + palbociclib (PALBO) + fulvestrant (FULV).
28 Background: INAVO is an oral, highly potent, and selective PI3Kα inhibitor that also promotes mut p110α degradation. In INAVO120 (NCT04191499), INAVO + PALBO + FULV showed statistically significant and clinically meaningful improvements in investigator-assessed progression-free survival (PFS; hazard ratio [HR] at primary analysis 0.43; 95% confidence interval [CI] = 0.32–0.59; p < 0.0001) and overall survival (OS; HR at final analysis 0.67; 95% CI = 0.48–0.94; p = 0.0190) v PBO + PALBO + FULV. We report subgroup analyses of Asian pts. Methods: Pts received INAVO (9 mg orally once daily [PO QD]) + PALBO (125 mg PO QD; Days 1–21 of each 28-day cycle) + FULV (500 mg intramuscularly; Cycle 1 Days 1 and 15 then every ~4 weeks) or PBO (PO QD) + PALBO + FULV. PFS, OS, duration of response (DoR), time to first subsequent chemotherapy (CT)/death, objective response rate (ORR), safety, and post-progression treatment (tx) were analyzed. Results: At clinical cut-off (Nov 15, 2024), 120 Asian pts were randomized (58 to INAVO; 62 to PBO). Median INAVO and PBO tx duration was 13.0 months (mo; range 0.9–52.3) v 6.2 mo (0.1–40.3), respectively. Median age was 52 years (range 27–77) in the INAVO arm v 50 years (33–79) in the PBO arm; 22.4 v 32.2% of pts had body mass index ≥25.0. Efficacy is shown in the Table. Grade 3–4 AEs occurred in 98.3 v 87.1% of pts; serious AEs, in 29.3 v 9.7% (most common: pyrexia [5.2%] and febrile neutropenia [5.2%] with INAVO v vomiting [3.2%] with PBO). AEs leading to INAVO/PBO discontinuation occurred in 3.4% (transitional cell carcinoma [1.7%], acute kidney injury [1.7%]) v 0% of pts; AEs leading to INAVO/PBO dose reduction, in 13.8 v 1.6% (most common: stomatitis [3.4%] v rash [1.6%]); AEs leading to INAVO/PBO dose interruption, in 72.4 v 27.4% (most common: hyperglycemia [31.0%] v neutrophil count decreased [8.1%]). The most common post-progression tx was CT (50.0 v 76.7% at second line). Conclusions: Consistent with the intention-to-treat (ITT) population, INAVO + PALBO + FULV showed a clinically meaningful efficacy benefit in Asian pts, with no new safety signals and numerically lower INAVO discontinuation and dose reduction rates v the ITT population. These data further support the benefit of the INAVO120 regimen in PIK3CA mut, hormone receptor-positive, HER2–, endocrine-resistant aBC. Clinical trial information: NCT04191499 . INAVO arm(n = 58) PBO arm(n = 62) HR (95% CI) Median PFS, mo (95% CI) 14.8 (9.3–21.1) 6.8 (5.4–9.2) 0.38 (0.24–0.59) Median OS, mo (95% CI) 32.7 (20.5–NR) 27.0 (20.1–NR) 0.74 (0.43–1.27) Median time to first subsequent CT/death, mo (95% CI) 19.2 (12.5–32.7) 10.6 (6.7–14.1) 0.47 (0.30–0.74) (n = 35) (n = 17) Median DoR, mo (95% CI) 18.8 (14.7–NR) 10.7 (7.5–NR) (n = 58) (n = 62) Δ, % (95% CI) ORR, % 60.3 27.4 32.9 (16.1, 49.7) NR, not reached.
Proteolysis‐Assisted Cyclization Facilitates Site‐Centric Target Deconvolution of Isothiocyanates
ABSTRACT Isothiocyanates (ITCs) are a unique class of electrophilic natural products that exert biological effects by reacting with proteinous cysteines to generate thionoacyl adducts. However, the identification of ITCs’ target sites is still an unmet task due to the high lability of such adducts. Here, we report an unexpected chemistry through which the ITC‐protein adduct forms a stable N‐terminal dihydrothiazole peptide adduct during proteolysis. This proteolysis‐assisted cyclization (PAC) reaction can be harnessed for developing affinity‐based and activity‐based chemoproteomic methods to site‐specifically profile targets of ITCs. Applying these methods not only adds further complexity to the known polypharmacological landscape of sulforaphane but also expands the ligandable cysteinome with site‐level resolution through a 55‐member ITC library. Given the promising chemopreventive and therapeutic effects of ITCs, the PAC‐based chemoproteomic platform may lay the groundwork for elucidating their mechanisms of action and ultimately diversifying cysteine targetability for drug discovery.
Allelopathic effects of Padina sp. and Sargassum sp. as biological control agents of harmful algal blooms (HABs) in the Persian Gulf and Gulf of Oman
Togetherness: How cooperation built the world
Over‐Strain‐Relaxation State and Dislocation‐Governed Polar Topologies in Ferroelectric Superlattices
ABSTRACT Polar topologies in ferroelectric films can be stably constructed by tensile strains through coherent growth. How potential polar textures in ferroelectric films grown under compressive strain substrate will evolve is still elusive. Here, we have grown high‐quality PbTiO 3 /SrTiO 3 superlattices on NdGaO 3 substrates, where the nominal mismatch of this system is about −1.0% and thus the superlattices may be subjected to compressive strain. Dislocation arrays between the substrate and the films were identified. Amazingly, these misfit dislocations were found to release the mismatch far exceeding the −1.0% nominal mismatch, approaching −1.7% and forming actually a final tensile strain state of about 0.7%. This strain state was driven by the competition between thermal mismatch dynamics and ferroelectric phase transition of PbTiO 3 , which triggers the formation of polar vortex arrays. In particular, the polar topology here has the tendency of transition from vortex to periodic dipole waves under the influence of threading dislocations. Our research provides possibilities of integrating and regulating ferroelectric topologies on nominal compressive strain substrates. The novel over‐strain‐relaxation behavior was unrealized previously, and will stimulate more strain manipulations for oxide epitaxial films. More efforts could be focused on some practical substrates, such as direct growth of polar topologies on silicon.
Ultraprecise Sign Language Recognition Realized by Self‐Recoverable Near‐Infrared Mechanoluminescent Materials
ABSTRACT Advancements in human‐machine interaction technology require flexible sensors to possess core capabilities, including high stability, anti‐interference properties, and self‐powering functionality. Traditional electrical sensors usually struggle to adapt to complex and long‐term application scenarios. Mechanoluminescence (ML) materials present a novel solution to this challenge, yet existing ML materials still suffer from issues such as requiring pre‐radiation charging and insufficient cycling stability. Here, we report a series of self‐recovering near‐infrared (NIR) ML materials—ZnGa 1‐ m Al m InO 4 :Cr 3+ , which possess excellent piezoelectric properties and low cost. By precisely controlling the crystal field strength through adjusting the doping concentration of Al 3+ ions, the photoluminescence intensity was enhanced by 40.65‐fold. Even after undergoing thousands of mechanical stimulation cycles, this self‐healing near‐infrared ML material retains 98% of its initial luminescence intensity. When integrated with photoelectric sensors, ZAIO:Cr 3+ @PDMS demonstrated outstanding performance in sign language recognition (achieving 99.46% accuracy) and intelligent road monitoring through convolutional neural networks. This work provides novel insights for designing NIR ML materials and lays the foundation for integrating ML materials with intelligent neural networks.