Multiscale Approximations to Understand the Complex Role of Microglia in Alzheimer's Disease Authors: Patricia Martínez‐Tazo, Marina Mejias‐Ortega, Verónica López‐López, Aysha M. Bhojwani‐Cabrera, Emma Saxton, Santiago Canals, Giles Hardingham, Antonia Gutierrez, Jose P. Lopez‐Atalaya, Silvia De Santis Published: 2026 ABSTRACT ABSTRACTMicroglia, the resident innate immune cells of the central nervous system, play a pivotal role in the pathogenesis of Alzheimer's disease (AD). Microglia are now recognized as a highly dynamic and heterogeneous population whose molecular and functional states vary with spatial context, disease stage, and genetic background. Recent discoveries across multiple scales from genetics, molecular and cellular biology, to systems‐level imaging and epidemiology have underscored the complex and context‐dependent contributions of microglia to the AD cascade. Together, these findings highlight the need for integrative, multiscale approaches that bridge molecular, cellular, and systemic perspectives to elucidate the diverse roles of microglia and their impact on disease progression. This mini‐review discusses recent advances in understanding microglial biology across these dimensions and outlines current challenges toward achieving a more unified and therapeutically oriented framework for studying microglia in AD. Abbreviations Alzheimer's disease apolipoprotein E apolipoprotein E ε4 allele amyloid‐β cerebral amyloid angiopathy cyclic GMP–AMP synthase cyclooxygenase–1 colony‐stimulating factor 1 receptor disease‐associated microglia diffusion‐weighted magnetic resonance imaging fluorodeoxyglucose forkhead box O3 hypoxia‐inducible factor 1 interferon/type I interferon interferon alpha/beta receptor 1 interleukin 1 alpha interferon‐responsive microglia lipid‐droplet–accumulating microglia lipopolysaccharide mean diffusivity microglia associated with neurodegeneration major histocompatibility complex class II magnetic resonance imaging mitochondrial DNA positron emission tomography prolyl hydroxylase domain‐containing protein 3 Spi1 transcription factor reactive oxygen species senescence‐associated secretory phenotype single‐cell RNA sequencing single‐nucleus RNA sequencing stimulator of interferon genes spleen tyrosine kinase terminally inflammatory microglia tumor necrosis factor triggering receptor expressed on myeloid cells 2 translocator protein white‐matter–associated microglia Introduction: Microglia in Alzheimer's Disease Alzheimer's disease (AD) is the leading cause of dementia, characterized clinically by progressive cognitive decline and neuropathologically by extracellular amyloid‐β (Aβ) plaques, intracellular neurofibrillary tangles composed of hyperphosphorylated tau protein, and widespread neurodegeneration (2025 Alzheimer's disease facts and Figures 2025). The amyloid cascade hypothesis, which posits that Aβ accumulation triggers downstream tau pathology, synaptic loss, and neuronal death, has long provided a conceptual framework for understanding AD pathogenesis (Hardy and Higgins 1992; Selkoe and Hardy 2016). However, it is now evident that this model alone cannot fully account for the complexity of the disease. Neuroinflammation, vascular alterations, aging‐related changes, and dysfunction of the glymphatic system act upstream of, or in parallel with Aβ accumulation, shaping the cellular landscape of AD (Hardy 2025). Within this multifactorial network, microglia, the resident innate immune cells of the central nervous system, emerge as central orchestrators of AD pathophysiology (Hansen et al. 2018; Leng and Edison 2021; Gao et al. 2023; Heneka et al. 2025; Chen et al. 2026). These highly dynamic cells continuously survey the brain microenvironment, remodeling synapses and responding to injury and protein aggregates through phagocytosis and cytokine secretion. Notably, genome‐wide association studies showed that a substantial proportion of genes susceptible for AD are expressed predominantly in microglial response pathways (Bellenguez et al. 2022), underscoring their pivotal role in disease susceptibility and progression. Experimental models further support this connection: in mouse models of amyloid pathology, the most prominent transcriptional response of microglia to Aβ aggregation is the upregulation of APOE, the gene encoding apolipoprotein E, the strongest genetic risk factor for late‐onset AD (Corder et al. 1993), the most common nonfamilial form of the disorder. Epidemiological studies have long hinted at a contribution of inflammation to AD, with early reports suggesting an inverse association between nonsteroidal anti‐inflammatory drug use and AD prevalence (In'T Veld et al. 2001). However, clinical trials with anti‐inflammatory interventions have largely failed to modify disease progression, pointing to the complex and context‐dependent roles of microglia in health and disease. Finally, the dissociation between amyloid burden and clinical outcome, observed in resilient individuals who remain cognitively intact despite high plaque load (Rowe et al. 2010), has attracted attention toward the modulatory role of microglia in amyloid and tau seeding. Indeed, the combined presence of Aβ, tau, and microglia shows a strong association with cognitive impairment (Pascoal et al. 2021), indicating that neuroinflammatory processes may influence the timing and progression of neurodegenerative vulnerability. In summary, microglia are unquestionably central to the pathophysiology of AD, exerting influence across multiple scales from molecular and cellular mechanisms to network‐level and even epidemiological dimensions. However, their precise contributions across disease stages remain only partially understood, reflecting the complexity and context‐dependence of their roles. In this framework, the present mini‐review does not aim to provide an exhaustive overview of microglial biology in AD, but rather to integrate recent findings across multiple scales, from molecular and cellular mechanisms to systems‐level dynamics, and from rodent models to human studies (Figure 1), as discussed in the recent symposium “Multiscale approximations to understand microglia dysfunction in AD: from genes to cells, looking for new therapeutic targets” of the Spanish Society for Neuroscience (SENC) meeting (Las Palmas, 3–5 September 2025). By synthesizing recent evidence across these dimensions, our goal is to provide a unified and translational perspective on how microglia influence the initiation, progression, and clinical manifestation of AD, and highlight key challenges and open questions that will shape future research. Overview of the multiscale and cross‐species approach adopted in this review. We integrate recent findings from molecular and transcriptomic studies, cellular and network‐level analyses, and systems or imaging approaches in both rodent models and humans, to provide a unified view of microglial contributions to Alzheimer's disease. Microglia Subtypes and Transcriptional States in Alzheimer's Disease Microglia comprise a heterogeneous population whose functions in AD evolve across distinct activation states. The traditional M1/M2 paradigm, which dichotomized microglial activation into pro‐ and anti‐inflammatory states, has been superseded by transcriptomic analyses revealing a continuum of highly plastic, context‐dependent phenotypes (Ransohoff 2016; Paolicelli et al. 2022). Single‐cell and spatial transcriptomic studies have demonstrated that microglia acquire distinct molecular profiles that shift with aging, the surrounding milieu, and disease‐related stimuli (Butovsky and Weiner 2018; Silvin and Ginhoux 2018; Stratoulias et al. 2019; Gerrits et al. 2020; Masuda et al. 2020; Chen and Colonna 2021; De Schepper et al. 2021; Provenzano et al. 2021; Paolicelli et al. 2022; Dadwal and Heneka 2024). In AD, single‐cell and spatial transcriptomic analyses have uncovered transcriptional programs that accompany amyloid deposition and disease progression, underscoring the role of microglia as both sensors and effectors of neurodegeneration (Keren‐Shaul et al. 2017; Mathys et al. 2019; Sala Frigerio et al. 2019; Olah et al. 2020; Prater et al. 2023; Sun et al. 2023; Tuddenham et al. 2024; Ardura‐Fabregat et al. 2025; Marshe et al. 2025). Among the microglial states identified through these approaches, disease‐associated microglia (DAM) are the most extensively characterized. Initially described in the 5xFAD amyloidosis model, DAM are defined by suppression of homeostatic genes such as P2ry12 and Tmem119 and induction of a transcriptional program enriched for phagocytic and lipid metabolism genes, including Apoe, Ctsd, Lpl, Tyrobp, and Trem2, several of which correspond to established genetic risk factors for AD (Keren‐Shaul et al. 2017; Krasemann et al. 2017; Bellenguez et al. 2022). This activation unfolds through a two‐step process comprising an early TREM2‐independent phase (DAM1) followed by a TREM2‐dependent transition (DAM2), enabling microglia to detect and respond to pathological stimuli such as amyloid‐β plaques and apoptotic neurons while undergoing the metabolic reprogramming required for sustained phagocytic activity (Wang et al. 2015; Keren‐Shaul et al. 2017; Deczkowska et al. 2018; Zhou et al. 2020). Functionally, DAM form a periplaque microglial barrier that promotes amyloid compaction, debris clearance, and restriction of neuritic damage, and TREM2 loss‐of‐function variants—associated with increased AD risk (Guerreiro et al. 2013; Jonsson et al. 2013)—impair these protective responses, resulting in reduced plaque containment and exacerbated neuritic dystrophy (Yuan et al. 2016; Leyns et al. 2019). Moreover, TREM2‐dependent DAM activation limits amyloid‐associated tau seeding and spreading, whereas both TREM2 deletion and microglial depletion markedly enhance tau propagation around plaques (Gratuze et al. 2021). However, sustained activation of the TREM2‐APOE axis can drive lipid overload and pro‐inflammatory stress, shifting microglia toward dysfunctional states, including the closely related microglia associated with neurodegeneration (MGnD) program, which shares key transcriptional features with DAM (Krasemann et al. 2017; Clayton et al. 2021). Importantly, the functional consequences of TREM2 signaling are highly context‐dependent and shaped by disease stage and genetic background. Notably, in the presence of ApoE4, TREM2 deletion paradoxically exacerbates tau‐mediated neurodegeneration, highlighting the coexistence of TREM2‐dependent and TREM2‐independent microglial programs with opposing effects in Alzheimer's pathology (Gratuze et al. 2023). Collectively, these findings position the TREM2‐APOE axis at the intersection of genetic susceptibility and context‐dependent microglial functional adaptation in AD. While the DAM phenotype is well characterized in murine models of neurodegeneration, its presence in human AD remains debated. Some studies have identified microglial populations with DAM‐like features in human brains, but without a clear overrepresentation in Alzheimer's patients compared to controls (Prater et al. 2023; Sun et al. 2023). In addition, most transcriptomic studies in post‐mortem human tissue use single nuclei sequencing (snRNA‐seq), a technique that, although more suitable for frozen samples, has lower transcriptional resolution than scRNA‐seq, especially regarding the detection of genes associated with active cytoplasmic functions (Thrupp et al. 2020). These methodological differences, together with the higher biological heterogeneity in humans, could explain the lack of direct correspondence between murine and human DAM profiles. While DAM largely reflect a lipid‐ and phagocytosis‐driven response to amyloid pathology, transcriptomic analyses have also revealed a second, orthogonal axis of microglial activation dominated by interferon signaling. This population has been termed interferon‐responsive microglia (IRM) (Sala Frigerio et al. 2019; Olah et al. 2020). IRM are defined by the induction of type‐I interferon (IFN‐I)‐stimulated genes, including Ifitm3, Isg15, Irf7, Oasl, and Stat1 (Lopez‐Atalaya and Bhojwani‐Cabrera 2025). In contrast to DAM, IRM exhibit a robust concordance between murine models and human patient samples (Sala Frigerio et al. 2019; Olah et al. 2020; Prater et al. 2023; Sun et al. 2023). In the human Alzheimer's cortex, microglia displaying this interferon‐driven profile become increasingly abundant with advancing pathology, consistent with a gradual reinforcement of innate immune and antiviral programs (Mathys et al. 2019). Functionally, these microglia show heightened interferon activity that promotes microgliosis, synaptic engulfment, and altered neuronal activity (Mathys et al. 2019; Roy et al. 2020). In vivo, blockade or genetic deletion of Ifnar1 attenuates microglial activation, rescues synaptic and cognitive deficits, and reduces amyloid‐associated inflammation (Roy et al. 2022). Experimental evidence indicates that sustained activation of interferon pathways in microglia is sufficient to trigger neurodegenerative processes independently of amyloid accumulation. In particular, deletion of Rela in microglia reprograms them into IRM, which leads to widespread neuroinflammation, axonal injury, and neuronal dysfunction despite a paradoxical reduction in amyloid burden (López‐López et al. 2026). The same study reported a pronounced enrichment of interferon‐related transcriptional signatures in female AD brains and in female APP/PS1 mice, linking sex‐biased interferon regulation to the heightened vulnerability of women to AD. Collectively, these findings demonstrate that aberrant interferon signaling within microglia can act as a pathogenic amplifier, coupling innate immune activation to neuronal stress and synaptic dysfunction, and highlight this pathway as a potential target for therapeutic intervention. Notably, the APOE3 Christchurch R136S mutation confers resilience against tau pathology through selective attenuation of microglial cGAS‐STING‐IFN‐I signaling, highlighting microglial innate immune pathways—particularly cGAS‐STING‐IFN‐I signalling—as promising therapeutic targets for AD and other tauopathies (Naguib et al. 2025). Additional microglial populations have been described in AD, including MHC‐II‐high antigen‐presenting microglia that may interact with infiltrating T cells (Olah et al. 2020); in the aged brain such as lipid‐droplet‐accumulating microglia (LDAMs), characterized by defective lipid metabolism and elevated oxidative stress (Marschallinger et al. 2020); and white‐matter‐associated microglia (WAMs), enriched for Lpl, Spp1, and Gpnmb and prevalent in aged white matter (Safaiyan et al. 2021). Although their contribution to disease progression remains less clearly defined, these phenotypes further illustrate the adaptive range of the microglial response in AD. More recently, a distinct microglial population has been identified in AD brains that accumulates in carriers of the APOE4 genetic variant (Millet et al. 2024). These cells, termed terminally inflammatory microglia (TIMs), are characterized by simultaneous expression of inflammatory signals and cellular stress markers, suggesting an exhausted‐like state similar to that observed in chronically stimulated T cells. TIMs exhibit functional impairments, including defective amyloid‐beta clearance capacity. These findings suggest that microglial exhaustion may represent a key mechanism underlying AD pathology and progression, particularly in high‐risk populations (Millet et al. 2024). Emerging evidence suggests that intrinsic microglial states defined by distinct transcriptional and immunoregulatory programs can bias these functional outcomes. For example, a recently described PU.1low CD28+ microglial subpopulation exhibits neuroprotective properties in AD models, limiting neuroinflammation and amyloid pathology, thereby highlighting how fine‐tuned transcriptional regulation may influence the balance between pathogenic and protective microglial functions (Ayata et al. 2025). Microglial activation in AD thus seems to encompass a continuum of transcriptional and functional states rather than a binary model. While certain phenotypes such as DAMs and IRMs exemplify recurrent and well‐characterized responses, they likely represent interconnected stages within a broader landscape of activation. Their balance—between lipid processing, interferon signaling, and immune reactivity—appears to evolve with disease stage and local environment. Defining the molecular checkpoints that govern transitions between transcriptional states will be critical for the development of interventions capable of fine‐tuning microglial function in a stage‐ and context‐specific manner. Friend or Foe? The Complex Role of Microglia in Alzheimer's Disease Plaque Biology Given their phagocytic capacity within the central nervous system, microglia were long anticipated to facilitate the clearance of extracellular amyloid‐β (Aβ) aggregates, thereby exerting a protective effect against AD pathology. However, recent studies have revealed a far more complex and context‐dependent role. Evidence support that microglia play a multifaceted and stage‐dependent role in Aβ plaque biology, acting not only as phagocytes but also as structural regulators of plaque architecture. Early in pathology, microglia respond to emerging Aβ aggregates by rapidly extending processes toward nascent deposits and forming a physical barrier that constrains plaque growth, a response driven by TREM2, ApoE, and SYK‐dependent signaling pathways that coordinate chemotaxis, metabolic reprogramming, and Aβ uptake (Condello et al. 2015; Wang et al. 2015; Yuan et al. 2016; Ulland et al. 2017). Loss of TREM2 function or expression of genetic variants, such as R47H, disrupts this coordinated response, resulting in decreased microglial clustering, impaired plaque compaction, and the emergence of more filamentous and neurotoxic plaques (Jay et al. 2015, 2017; Cheng‐Hathaway et al. 2018; Wood et al. 2022; Liu et al. 2023). ApoE, which is strongly upregulated in plaque‐associated microglia, further modulates lipid metabolism and Aβ aggregation (Keren‐Shaul et al. 2017; Krasemann et al. 2017). Its absence or replacement with ApoE4 impairs microglial encapsulation of plaques and aggravates neuritic pathology (Ulrich et al. 2018; Liu et al. 2023). Downstream of these receptors, SYK integrates signals from TREM2, CLEC7A, and TAM receptors to drive the transition toward a DAM phenotype that supports Aβ phagocytosis and plaque compaction; microglial Syk deletion therefore increases overall plaque burden and shifts plaque morphology toward more diffuse or dynamic structures (Ennerfelt et al. 2022; Wang et al. 2022). Paradoxically, genetic and pharmacological microglial depletion studies have revealed that microglia also promote plaque initiation. While microglia were long presumed to act primarily as Aβ‐clearance cells, CSF1R inhibitor–based ablation studies (Sosna et al. 2018; Spangenberg et al. 2019) and microglia‐deficient mouse models such as CSF1RΔFIRE/ΔFIRE (Rojo et al. 2019) demonstrate that early removal of microglia can reduce the number of plaques. In CSF1RΔFIRE/ΔFIRE mice crossed onto models of ß‐amyloidopathy, microglial absence markedly decreased neocortical plaque density (Baligács et al. 2024; Li et al. 2024), and predeposition depletion similarly reduced plaque numbers and neuritic pathology (Baligács et al. 2024), supporting a role for microglia in plaque seeding. Notably, when the CSF1RΔFIRE/ΔFIRE mouse was crossed with the aggressive 5xFAD amyloidopathy model, the resulting animals developed severe cerebral amyloid angiopathy, intracerebral hemorrhages, and premature lethality (Kiani Shabestari et al. 2022). This outcome indicates that the impact of microglial loss is highly dependent on genetic background and disease severity, reinforcing the notion that microglial functions in AD are strain‐ and model‐specific. Notably, these studies indicate that while microglia facilitate the earliest stages of plaque formation, plaque core maturation may be less dependent on microglial presence, as remaining plaques still display dense methoxy‐04–positive cores (Li et al. 2024). At later stages, however, microglia acquire a protective function and their removal leads to diminished plaque compaction, increased fibrillar Aβ, and worsening neuritic degeneration, consistent with mechanistic work implicating microglial encapsulation as a barrier against toxic Aβ species. Despite reducing parenchymal amyloid, microglial depletion frequently exacerbates cerebral amyloid angiopathy, indicating that microglia also influence the spatial routing of Aβ deposition. Human–microglia chimeric mouse models further highlight the importance of microglial activation states and genetic background, demonstrating that microglial responsiveness promotes both early plaque containment and later plaque compaction, even as chronic inflammatory activation may contribute to neuronal vulnerability (Baligács et al. 2024). For example, microglia carrying the TREM2^R47H/R47H mutation exacerbate amyloid plaque pathology, highlighting the critical role of properly responsive (rather than hypo‐reactive) microglia in restraining early events of the amyloid cascade (Baligács et al. 2024). Taken together, these observations indicate that the contribution of microglia to Aβ pathology is highly complex, stage‐ and context‐dependent, varying according to disease progression, cerebral region, genetic background, and experimental conditions. One possible mechanistic model proposes that microglia actively take up Aβ but fail to fully degrade it, leading to intracellular accumulation of partially processed Aβ. This material may subsequently be released under cellular stress, acting as a seed for extracellular amyloid deposition (Kim et al. 2012; Ulrich et al. 2018; Bassil et al. 2021; Kaji et al. 2024). Further complicating our understanding of the role of microglia in AD is their extensive crosstalk with other central nervous system cell types, especially astrocytes, as both glial populations are integral elements of the innate immune system based on their ability to produce immunomodulators and expression of receptors associated with innate immunity. Emerging single‐cell and spatial transcriptomic data reveal that microglia–astroglia interactions are highly context‐dependent and concentrated in the amyloid plaque niche, where specialized plaque‐associated glial states and signaling gradients reshape local inflammation and synaptic vulnerability (Mallach et al. 2024). Microglia can induce astrocyte reactivity by secreting inflammatory cytokines IL‐1α, TNF, and C1q (Liddelow et al. 2017). In addition, they modulate Aβ effects on astroglial reactivity (Ferrari‐Souza et al. 2026). These reactive astrocytes exhibit a dysfunctional neurotoxic state with loss of neurotrophic functions (Guttenplan et al. 2021; Patani et al. 2023). In turn, astrocyte‐derived inflammatory cytokines and chemokines modulate microglial phagocytic function (Lian et al. 2016), thus creating self‐reinforcing inflammatory circuits that exacerbate Aβ pathology. Genetic risk factors, like TREM2 variants and APOE4, further tune this crosstalk impairing amyloid clearance and promote maladaptive glial activation in AD (Yin et al. 2023). Finally, like microglia, astrocytes too can adopt neuroprotective as well as harmful responses in AD models (Jiwaji et al. 2022). Beyond active neurotoxicity of reactive astrocytes, they are also known to lose many of their neuro‐supportive homeostatic functions (Fuller et al. 2010; Escartin et al. 2021; Jiwaji and Hardingham 2022). These dynamic glial interactions underscore that microglia should be studied not as isolated entities but rather as components of a coordinated, multicellular network governing brain homeostasis, innate immunity, and neurodegeneration. Moreover, microglia‐astroglia dialog could become an effective target for future AD treatment. The current challenge lies in translating these mechanistic insights into therapeutic strategies. While recent studies have illuminated the multifaceted roles of microglia in both initiating and remodeling amyloid pathology, it remains unclear how to modulate microglial activity to enhance their protective functions without triggering chronic inflammation or neurotoxicity. This question has become particularly pressing in the era of anti‐amyloid antibody therapies, which effectively clear plaques but leave open the issue of how the brain's innate immune system, especially microglia, responds to and participates in this clearance process. For example, recent evidence underscores the need to carefully consider the timing and context of AD therapeutic treatments, especially in the cases of therapies that directly or indiscriminately suppress the activation of microglia (Cadiz et al. 2024). Integrating our growing molecular and functional understanding of microglia with emerging immunotherapeutic approaches will be essential for developing next‐generation interventions that harness microglial biology to promote resilience rather than degeneration in AD. Alzheimer's Disease: A Human‐Only Pathology AD is a human‐specific neurodegenerative condition, and understanding its pathophysiology requires moving research beyond rodent models. While transgenic mouse models harboring familial AD‐associated mutations have been instrumental in elucidating molecular and cellular mechanisms, they only partially recapitulate the complex human pathology (Sanchez‐Varo et al. 2022; Trujillo‐Estrada et al. 2022). In this sense, innate immune responses differ substantially between AD patients and transgenic mice, with human AD microglia showing progressive functional decline, metabolic and transcriptional exhaustion, and dystrophic morphology in late‐disease stage, whereas rodent microglia typically mount a robust and sustained microglial activation in response to amyloid deposition (Sanchez‐Mejias et al. 2016; Navarro et al. 2018). Moreover, amyloid seeds derived from human AD brains induce a more aggressive and extensive amyloid pathology than those obtained from aged transgenic mice (Andreo‐Lopez et al. 2024), underscoring fundamental interspecies differences in amyloid structure, propagation, and microenvironmental interactions. Advanced platforms, such as transgenic mice harboring xenografted human microglia (Mancuso et al. 2024; Murphy et al. 2025), have revealed valuable insights into human‐specific transcriptional responses and interactions with amyloid deposits; however, they still reproduce only a limited fraction of the human disease landscape, lacking key features including tau aggregation, vascular pathology, systemic influences, and the genetic diversity inherent to human populations. These differences have significant translational implications, as therapeutic strategies that appear effective in rodent models fail to produce clinical benefits in patients due to the incomplete representation of human disease mechanisms, thereby limiting predictive value for clinical outcomes. Consequently, there is a critical need for complementary experimental systems that more faithfully model human pathology. Achieving this requires first elucidating microglial pathology and behavior directly within the human AD brains. Histopathological analyses of postmortem AD brains reveal that microglia often display a dystrophic morphology, characterized by fragmented, beaded, and deramified processes, spheroidal swellings, and cytoplasmic cytorrhexis (Streit et al. 2009; Sanchez‐Mejias et al. 2016; Navarro et al. 2018; Swanson et al. 2020). Unlike activated microglia, dystrophic microglia display impaired functional capacity leading to a decline in the brain's innate immune surveillance and protein clearance mechanisms, contributing to impaired amyloid clearance, synaptic dysfunction, and progression of neurodegeneration in AD. Dystrophic microglia are significantly more abundant in early‐affected brain regions and correlate with the spread of tau and amyloid pathology, suggesting a disease‐specific, possibly causative role in neurodegeneration (Shahidehpour et al. 2025). Beyond morphology, transcriptomic evidence points to an exhausted microglial state in human AD, with downregulation of homeostatic, metabolic, and immune‐regulatory pathways, consistent with functional senescence (Millet et al. 2024). Emerging evidence supports microglial senescence in AD (Rim et al. 2024), characterized by cell cycle arrest, morphological dystrophy, and the secretion of proinflammatory mediators, a profile known as the senescence‐associated secretory phenotype (SASP). In the AD context, this SASP not only sustains chronic neuroinflammation but also disrupts neuronal homeostasis and weakens microglial ability to clear amyloid‐β, thereby promoting plaque accumulation (Millet et al. 2024). Therefore, a deterioration of microglial function, rather than microglial activation, seems to contribute to AD development in humans. Mitochondrial dysfunction may be a central driver of the maladaptive microglial state observed in AD brains. Because effective microglial activation depends heavily on oxidative phosphorylation (March‐Diaz et al. 2021), under conditions of brain hypoxia resulting from ischemia, reduced perfusion, or chronic vascular insufficiency (age‐associated risk factors), microglia fail to fully activate and are less capable of migrating and clustering around amyloid plaques. This insufficient recruitment compromises the formation of the protective microglial barrier, thereby facilitating further neurotoxic spread. In addition, impaired mitochondria generate excessive reactive oxygen species (ROS) and release mitochondrial DNA (mtDNA) into the cytosol, that together with age‐related iron accumulation (Kenkhuis et al. 2021), amplify oxidative stress and render microglia increasingly vulnerable to ferroptosis, an iron‐dependent lipid peroxidation–mediated form of cell death implicated in AD pathology (Qiu et al. 2023). Ferroptotic signaling further destabilizes cellular membranes, exacerbates inflammatory responses, and undermines microglial homeostatic functions. In parallel, cytosolic mtDNA serves as a potent activator of the cGAS–STING pathway as demonstrated in mice (Gulen et al. 2023), eliciting type‐I interferon responses that reinforce both cellular senescence and chronic inflammatory activation (Jin et al. 2022; Zhang et al. 2024; Lopez‐Atalaya and Bhojwani‐Cabrera 2025). In addition, a recent translational study showed that HIF1‐driven induction of PHD3 in amyloid‐associated microglia activates a FOXO3‐mediated interferon response, which impairs microglial function and amplifies AD pathology (Sanchez‐Garcia et al. 2025). Collectively, these processes establish a self‐perpetuating pathological loop linking mitochondrial impairment, oxidative injury, ferroptosis, type‐I interferon response, and senescence that accelerates microglial dysfunction and contributes to neurodegeneration (Rim et al. 2024). Therapeutically, interventions aimed at mitigating senescence, stabilizing mitochondrial health, reducing oxidative and ferroptotic stress, or modulating maladaptive interferon signaling through the PHD3‐FOXO3 axis or cGAS‐STING pathway hold substantial promise for restoring microglial resilience in AD. This AD‐associated dysfunctional state of microglia may promote the recruitment of peripheral macrophages into the diseased brain, potentially as a compensatory response to impaired immune surveillance. However, definitively demonstrating such recruitment has long been challenging, largely because of the historical lack of specific markers capable of reliably distinguishing resident microglia from infiltrating macrophages in human tissue. In this sense, combining immunohistochemistry with single‐cell transcriptomic markers has recently revealed the infiltration of monocyte‐derived myeloid cells into the hippocampus of human AD brains, where they integrate into both parenchyma and amyloid plaques (Muñoz‐Castro et al. 2023). Notably, in approximately 60% of Braak stage V–VI individuals with dementia, these infiltrating cells expressed genes rarely found in resident microglia but abundant in peripheral monocytes, challenging the long‐held assumption that all plaque‐associated myeloid cells are microglial in origin. In addition, distinct amyloid‐β and tau–associated microglial profiles have been identified in human AD tissue (Gerrits et al. 2021), further underscoring the heterogeneous and dynamic immune landscape of the human brain in AD. Complementary genetic evidence further supports this complexity: An isoform varying genetic modification in the microglial SIRPβ1 receptor coding gene has been shown to modulate microglial activation states and exert a dual effect on cognitive decline, linking human genetic diversity directly to functional and clinical variability in AD (García‐Alberca et al. 2024). Despite these advances, most human data still derive from post‐mortem analyses, which are limited by small sample sizes, variable control cohorts, and inconsistent methodologies across studies. To fully capture the temporal and spatial dynamics of microglial responses in living humans, the field urgently needs standardized multicenter protocols, larger datasets, and the development of noninvasive in vivo biomarkers capable of monitoring microglial activity longitudinally. Bridging this translational gap is essential to design therapies that are not only effective in models but also relevant to human disease biology. Aiming at a More Systemic and Translational View: In Vivo Microglia Imaging If the role of microglia in AD is stage‐ and context‐dependent, understanding their contribution in humans across different pathological stages becomes essential. However, direct access to such information remains extremely limited, as postmortem analyses provide only static snapshots of the disease. Therefore, in vivo imaging approaches capable of probing microglial activity are indispensable for capturing the dynamic nature of microglial responses during disease progression. Positron emission tomography (PET) represents the most widely used technique for in vivo molecular imaging of neuroinflammation in AD, allowing the detection of proteins upregulated in reactive microglia and astrocytes (Calsolaro and Edison 2016; Harput et al. 2026). Progressive increase in binding of the translocator protein (TSPO), a stress‐ and activation‐responsive mitochondrial protein whose expression is induced by pro‐inflammatory signals, has been proposed as a biomarker of disease evolution (Rupprecht et al. 2010; Kreisl et al. 2016). Thanks to multimodal labeling, several studies have suggested that microglial reactivity occurs at very early stages in the disease (Hamelin et al. 2016) and mediates relationships between different pathological manifestations (Rossano et al. 2024). Indeed, in individuals with mild cognitive impairment due to early‐onset AD, increased microglial reactivity has been detected even in the absence of amyloid tracer and prior to significant tau accumulation (Okello et al. 2009; Kreisl et al. 2016), supporting the view that neuroinflammation represents an early event in the pathological cascade. Moreover, PET imaging has enabled the spatial association between microglial reactivity and tau propagation following Braak stages (Pascoal et al. 2021). Despite these substantial findings, some limitations of PET‐based neuroinflammation imaging remain under debate. In particular, the interpretation of TSPO binding as a direct marker of microglial reactivity has been questioned. While studies in murine models support this association (Pascoal et al. 2021), recent evidence in humans suggests that TSPO signal may rather reflect the regional density of microglia than their reactivity state (Nutma et al. 2023), raising concerns regarding the translational validity of this tracer. To this end, several emerging PET tracers targeting microglia‐related pathways, such as CSF1 receptors (Horti et al. 2019; Zhou, Ji, Seki, et al. 2021), P2X7 receptors (Beaino et al. 2021; Moreno et al. 2026), P2Y12 receptors (Joseph et al. 2025 ), and COX‐1 (Zhou, Ji, Kong, et al. 2021), are being developed to provide greater cellular specificity and hold promise for improving the characterization of neuroinflammatory processes in AD. Recent work investigating the cellular contribution to other commonly used PET tracers in AD has highlighted a potential role of microglia in metabolic imaging signals. Although FDG‐PET is classically interpreted as a marker of neuronal and synaptic dysfunction associated with neurodegeneration, and a potential early tracer of the disease (Iaccarino et al. 2019), both mice and human studies have shown that microglia display a higher glucose metabolic rate than neurons and astrocytes (Xiang et al. 2021). These findings suggest that microglial state contributes to the observed alterations in FDG‐PET signal seen in mouse models of neurodegenerative diseases, and thus neuroimmune‐metabolic interactions may significantly influence the interpretation of PET signals in AD. In addition to the aforementioned limitations in specificity, PET also faces constraints related to spatial resolution (Cumbers et al. 2024) and relies on ionizing radiation, which may limit its applicability in vulnerable populations. Therefore, alongside molecular and cellular approaches, there is a need for system‐level, translational methods capable of probing microglial function in the living brain. Magnetic resonance imaging (MRI) can complement PET and provide such a bridge. MRI enables the noninvasive visualization of multiple aspects of brain anatomy and physiology without the use of ionizing radiation or exogenous contrast agents. Its translational nature allows identical imaging modalities to be applied across species, thereby bridging findings from animal models to human patients (De Santis et al. 2019). In AD, volumetric MRI remains one of the most widely used approaches, as neurodegeneration is closely associated with progressive brain atrophy. Classic longitudinal studies have demonstrated a dynamically spreading wave of grey matter loss across cortical regions during disease progression (Thompson et al. 2003). More recent work has extended these observations to preclinical and prodromal stages, showing significant grey matter volume loss in individuals with subjective cognitive decline, particularly within regions typically vulnerable to AD pathology (Arrondo et al. 2022). Beyond volumetric analyses, microstructural MRI based on diffusion‐weighted (dw) signals enables the probing of cellular‐scale features of brain tissue architecture. Initially developed to study the organization of white matter axons, dw‐MRI has since been applied to grey matter, where metrics such as mean diffusivity (MD)—the average rate of water diffusion—have been used to infer microstructural complexity and integrity (Ganepola et al. 2018; De Santis et al. 2020). In AD, increased MD is generally interpreted as a loss of microstructural organization (Weston et al. 2015). Recent large‐scale human studies have confirmed the sensitivity of MD to early pathological processes. A longitudinal study of nondemented individuals stratified by Aβ/tau status reported widespread cortical MD increases over time, particularly in Aβ‐positive/tau‐negative participants, with these changes correlating with astrocytic biomarkers and cognitive decline (Spotorno et al. 2024). A recent commentary on this paper further highlighted the potential of cortical MD as a sensitive marker of early microstructural alterations in AD, capturing changes that may reflect glial responses and neuroinflammation beyond conventional volumetric measures (Rodriguez‐Vieitez et al. 2024). Another study demonstrated that cortical MD is more sensitive than cortical thickness in detecting AD‐related structural alterations (Sun et al. 2024). Furthermore, a systematic review identified consistent associations between AD genetic risk and increased MD across temporal and frontal cortices, the cingulum, and corpus callosum (Harrison et al. 2020). Interestingly, presymptomatic carriers of familial autosomal‐dominant Alzheimer disease mutation show an inverse initial pattern reporting a reduced cortical MD, subsequently increased with symptom onset, showing a steeper rise than noncarriers (Vilaplana et al. 2020). The same study found an inverse correlation between MD and astrocytosis. When crossed with neuropathological data, MD demonstrated sensitivity to pathological biomarkers linked to neuritic plaque density, Braak stage and vascular health (Alasmar et al. 2026). While neurons constitute the largest volumetric contribution to the MRI signal in grey matter, glial cells occupy a nonnegligible and dynamic fraction of the parenchyma. Given their ability to undergo rapid morphological remodeling, migration, and proliferation, glial populations—particularly microglia—are likely to contribute to the observed contrast in diffusion metrics. Indeed, we have found that regional variability in MD across the mouse brain correlated more strongly with glial than neuronal density (Martínez‐Tazo et al. 2024), emphasizing the relevance of glial architecture to MRI‐derived microstructural measures. Beyond simple tensor‐based models, advanced diffusion frameworks have improved sensitivity to neurite orientation dispersion and density, as well as to soma size and density (Zhang et al. 2012; Palombo et al. 2020). Although these models inherently reflect contributions from microglia, they do not explicitly isolate them. However, we recently demonstrated that MRI can capture a distinct in vivo fingerprint of microglial morphological alterations following acute inflammatory challenges (Garcia‐Hernandez et al. 2022). This finding was subsequently replicated under systemic, more physiological inflammatory challenges (Kim et al. 2023), confirming that diffusion MRI metrics are sensitive to microglial reactivity. Moreover, diffusion‐based measures were shown to correlate with circulating lipopolysaccharide concentrations, suggesting their potential as noninvasive biomarkers of neuroimmune status (Birg et al. 2024). The major challenge ahead is both technical and translational. Although advanced diffusion MRI methods show promise in detecting microstructural signatures linked to neuronal degeneration and inflammation, their complex modelling frameworks and acquisition requirements currently limit their clinical penetrance. Bridging this gap will require the development of standardized acquisition protocols, robust cross‐site harmonization, and computational strategies capable of handling high‐dimensional imaging data. In this context, artificial intelligence‐based approaches have recently been proposed as powerful tools to extract biologically meaningful features from complex MRI datasets and improve diagnostic accuracy (Eggl and De Santis 2026; Jallais and Palombo 2024). These methods may ultimately help translate sophisticated microstructural models into clinically viable biomarkers of glial activity and neuroinflammation in AD. Conclusions Progress in understanding the role of microglia in AD has been remarkable, yet the field remains fragmented across molecular, cellular, and systems‐level approaches. Moving forward, greater integration across subfields and experimental scales from transcriptomics and imaging to genetics and clinical biomarkers will be essential to build a unified framework of microglial function in AD. Bridging data from rodent models, humanized systems, and in vivo human studies will help reconcile species differences and identify the mechanisms most relevant to human pathology and treatment. Ultimately, such cross‐disciplinary integration is key to transforming mechanistic insights into therapeutic strategies that are both biologically grounded and clinically effective. Author Contributions Patricia Martínez‐Tazo: investigation, writing – original draft, data curation, methodology, visualization, writing – review and editing. Marina Mejias‐Ortega: investigation, methodology, writing – original draft, writing – review and editing. Verónica López‐López: investigation, writing – original draft, methodology, writing – review and editing. Aysha M. Bhojwani‐Cabrera: investigation, writing – original draft, methodology, writing – review and editing. Emma Saxton: investigation, writing – original draft, methodology, writing – review and editing. Santiago Canals: conceptualization, writing – original draft, writing – review and editing. Giles Hardingham: conceptualization, investigation, writing – original draft, writing – review and editing, methodology, data curation, supervision, resources, funding acquisition. Antonia Gutierrez: conceptualization, investigation, methodology, validation, visualization, data curation, supervision, resources, funding acquisition, writing – original draft preparation, writing – review and editing. Jose P. Lopez‐Atalaya: conceptualization, investigation, funding acquisition, writing – original draft, methodology, validation, visualization, writing – review and editing, data curation, supervision, resources. Silvia De Santis: conceptualization, investigation, funding acquisition, writing – original draft, methodology, validation, visualization, writing – review and editing, data curation, supervision, resources. Funding This work was supported by the Instituto de Salud Carlos III (PI24/00274), the Generalitat Valenciana (PROMETEO 2020/007, CIESGT/2025/06, CIPROM/2022/15, CIDEGENT/2021/015), the Ministerio de Ciencia e Innovación (PID2021‐129053OB‐I00, PID2021‐128909NA‐I00, CNS2023‐144883, RTI2018‐102260‐B‐I00, PID2021‐128158NB‐C21, PID2024‐162400OB‐C21, PCI2024‐153491), the Fundacion Pasqual Maragall (2023–1296), the UK Dementia Research Institute, the Program for Centres of Excellence in R&D Severo Ochoa (CEX2021‐001165‐S) the Junta de Andalucía (CTS950‐G‐FEDER), and PMT is supported by the Spanish Ministry through a PhD fellowship FPU20/05737. Conflicts of Interest The authors declare no conflicts of interest.