The unique role of anosognosia in the clinical progression of Alzheimer’s disease: a disorder-network perspective

Abstract

Alzheimer’s disease (AD) encompasses a long continuum from a preclinical phase, characterized by neuropathological alterations albeit normal cognition, to a symptomatic phase, marked by its clinical manifestations. Yet, the neural mechanisms responsible for cognitive decline in AD patients remain poorly understood. Here, we posit that anosognosia, emerging from an error-monitoring failure due to early amyloid-β deposits in the posterior cingulate cortex, plays a causal role in the clinical progression of AD by preventing patients from being aware of their deficits and implementing strategies to cope with their difficulties, thus fostering a vicious circle of cognitive decline.

This article advances the view of anosognosia as a driver of cognitive decline in Alzheimer’s disease, due to early amyloid-β deposits in the posterior cingulate cortex, with error-monitoring failure and patients’ inability to counter their deficits.

Introduction

Alzheimer’s disease (AD), the leading cause of dementia, has become a relevant societal issue worldwide^1,2. Major advances in genetic analysis have made it possible to identify genes that contribute to an increased risk of developing the disease, often after the age of 65. This has helped distinguish between the sporadic, multifactorial, polygenic and highly prevalent form of AD, for which age is the main risk factor, and the rare, monogenic, and dominantly inherited form of AD, linked to mutations in three specific genes (APP, PSEN1, PSEN2) that are responsible for the manifestation of the disease much earlier in life^3,4. Current knowledge on sporadic AD points to pathophysiological changes that begin around 20 years before the clinical presentation of the disease, which may, in part, reflect the cumulative effect of different risk and protective factors throughout life^5,6. In fact, AD begins with a long asymptomatic or preclinical phase, characterized by specific neuropathological changes in the brain. The symptomatic phase ranges from a prodromal, predementia phase, with largely preserved autonomy, to a dementia phase with gradual loss of autonomy⁷.

Thanks to sensitive measures for detecting cognitive decline and pathological biomarkers associated with AD^8–10, it is now possible to monitor the disease’s evolution along its continuum^11,12. The cognitive decline that characterizes the symptomatic AD phase is expected to occur only in the presence of amyloid-β plaques and intracellular tau tangles—the hallmark pathological features of AD^13–15, along with synaptic dysfunction and/or neuronal loss.

Typically, prodromal AD patients present a relatively isolated amnestic syndrome of hippocampal type, characterized by episodic memory deficits, of which they are often unaware¹⁶. This intriguing phenomenon, known as anosognosia, can delay AD diagnosis and therapeutic interventions, compromise patient safety and increase the burden of caregivers¹⁷, leading to broader social impacts.

The term anosognosia—from the Greek words: a (without), nosos (disease), and gnosis (knowledge)—was introduced in 1914 by the French neurologist Joseph Babinski^18,19 to describe two patients who presented left hemiplegia secondary to a right hemisphere stroke, but manifested lack of awareness for their motor deficit (i.e., anosognosia for hemiplegia, AHP²⁰). Since then, anosognosia has been identified in several neurological conditions, with consistent evidence that it can affect motor²¹, sensory^22,23, behavioural and cognitive deficits^16,24–26, among others.

In current literature, the prevailing view is that anosognosia is domain-specific, which has given rise to several modular explanatory models closely dependent on the neurological condition and the affected domains^27–31. In contrast to this view, our recent model argued for a universal explanation of the neural mechanistic cascade underlying anosognosia for any type of deficit, regardless of the neurological condition³². The dual-path model is grounded on the notion that anosognosia is error-prone, meaning that it manifests only when errors are committed (like the error of forgetting in the context of memory deficits; or the inability to walk in the context of hemiplegia), which would explain why it can affect one domain more than another depending on the neurological condition and the level of severity of the related deficits. Following this concept, the emergence of anosognosia in the human brain would necessarily depend on a synaptic failure in the error-monitoring system resulting from either direct (i.e., damaged module) or indirect dysfunction (i.e., disconnected module) of key-related structures, such as the cingulate cortex, the orbitofrontal cortex (OFC) or the amygdala³². Notably, in addition to a direct path to anosognosia, resulting from damage to the cingulate cortex and consequent failure of the error-monitoring system, the disconnected module would contribute to the multifaceted spectrum of anosognosia manifestations displayed by the patients. For instance, an indirect dysfunction resulting from a deficit in emotional processing, in which the OFC and amygdala play a major role, would prevent patients from monitoring the internal milieu for relevant errors and assigning them an appropriate value, thus biasing their impact while evaluating one’s own performance. In such cases, patients could be relatively aware of their deficits, but quite indifferent to their consequences.

The brain’s synaptic function can be derived non-invasively from event-related potentials (ERPs). Two ERPs have been identified via electroencephalography (EEG) to assess error monitoring: the ERN, error-related negativity, and the Pe, positivity error³³. The ERN is response-locked and reaches a frontocentral negative maximum around 50–150 ms after the execution of an erroneous response, being presumably generated in the anterior cingulate cortex (ACC)³⁴, despite some evidence for maximal activity in the posterior cingulate cortex (PCC)³⁵. It is followed by a positive centroparietal deflection, the Pe, around 150–550 ms after the error occurrence, with the PCC as one of its main neural generators^34,36. Of note, whereas the ERN is considered to be a more automatic and preconscious process, the Pe has been critically associated with error awareness and appears to reflect more confidence in decision-making, with impact on learning. In particular, it has been suggested that an increase in Pe amplitude may reflect a learning strategy that allows subjects to avoid future errors by being particularly sensitive to previous ones^36–38. Accordingly, research carried out on non-human primates has confirmed that PCC activity increases its firing rate following the execution of an error and directly demonstrated that reversible inactivation of the PCC prevents new learning³⁹.

This has also been found in human subjects through patient studies. Notably, we observed opposite longitudinal effects on Pe amplitude during an episodic memory recognition task in two groups of amyloid-positive elderly subjects who were cognitively normal at baseline but differed in that one of them progressed to AD⁴⁰. Specifically, intra-group analyses revealed a significant reduction in Pe amplitude in the group of subjects who progressed to AD at the time of diagnosis, accompanied by decreased accuracies and slower response times. In contrast, the group that remained cognitively normal exhibited an increase in the Pe amplitude, coupled with improved accuracies and faster responses over the five-year study period. These contrasting Pe profiles were also observed in inter-group analyses, at the moment of diagnosis for those who progressed to AD, thus revealing a lack of awareness of memory impairment (that is, reduced awareness for erroneous responses in the memory task) in this group of subjects; versus a learning effect in those who maintained cognitive normalcy. This learning effect probably reflects an improved ability to recognize erroneous responses and adjust behaviour with task repetition⁴⁰. Nonetheless, at baseline, the group of subjects who progressed to AD had higher Pe amplitudes, while still cognitively normal, compared to the group of subjects who remained cognitively normal throughout the study períod, with the two groups exhibiting similar accuracy rates in the episodic memory task at that time point⁴⁰. This finding is of great relevance, as it suggests a preclinical less-wiring-more-firing neural compensatory mechanism to remedy the initial dysfunction of the error-monitoring system, occurring early in the AD continuum⁴¹. Overall, such ERP findings are consistent with the view that a synaptic dysfunction in the error-monitoring system might be present from the preclinical phase of AD, in agreement with previous work indicating that synaptic failure is a major determinant of this neurodegenerative disorder^42–44, and the best neuropathological correlate of cognitive decline⁴⁵. The role of the error-monitoring system has been confirmed by neuroimaging studies linking anosognosia to AD pathophysiology of the PCC^46,47 and suggesting PCC’s hypometabolism-induced anosognosia may predict the evolution to advanced stages of the disease⁴⁸.

Building up on these findings, the current perspective takes a different stance. Rather than viewing anosognosia merely as a risk factor or an independent predictor of AD progression, we propose that anosognosia might play a unique, causal role in the clinical decline over the AD continuum, with the PCC being its critical lesion site, and the central hub of a large-scale AD-related network.

Specifically, we formulate two main predictions:

• (i)A failure in the error-monitoring system would be at the core of self-awareness changes in the AD continuum, with a period of heightened awareness (hypernosognosia), in the preclinical phase, during which patients underestimate their abilities, although still cognitively normal, followed by a period of progressive loss of awareness (anosognosia), with overestimation of their cognitive abilities, which often characterizes the entrance in the clinical AD phase (Fig. 1). A dysfunction in the PCC—resulting from synapse alterations induced by amyloid-β deposits in this specific brain region—would be at the root of this process through a dysfunction in the error-monitoring system from the preclinical phase of this neurodegenerative disorder.
• (ii)Unawareness of errors committed in the context of cognitive deficits (anosognosia), such as forgetfulness in the context of memory deficits, would exacerbate these deficits due to the subjects’ inability to adopt learning strategies to cope with, or at least alleviate, their difficulties. Anosognosia would also disrupt the conscious retrieval of episodic memories^49,50, progressively worsening cognitive decline and ultimately leading to a vicious circle mechanism fostering the clinical deterioration of Alzheimer’s patients.

Fig. 1. Awareness changes along the Alzheimer’s disease continuum.

a At the preclinical phase, an increase in Pe amplitude occurs from the early amyloid-β deposits in the PCC, as a neural compensatory mechanism of less-wiring-more-firing type, leading to a more intense error-monitoring process with potential heightened awareness (hypernosognosia), helping subjects to still learn with their errors, adjust behaviour, and perform cognitive tasks accurately, as cognitively healthy subjects. b This compensatory mechanism becomes gradually defective, mainly due to the accumulation of amyloid-β in the PCC, with an increasing failure of the error-monitoring system at the synaptic level, leading to progressive loss of awareness (anosognosia), reduction in Pe amplitude, and consequent subjects’ inability to learn with their errors and implement strategies to counter their deficits.

It should be noted that amyloid-β burden has been associated with a progressive decline in self-awareness, reaching anosognosia about three years before the dementia stage⁵¹. The period of heightened awareness in the preclinical phase could probably reflect a more intense error-monitoring process, with a potential increase in Pe amplitude, enabling subjects to counter their emerging difficulties. Consistent with this, the Pe amplitude has been shown to be modulated by the individual significance of errors in the context of an emotional and motivational appraisal process^52,53. With the accumulation of amyloid-β and the progression of neurodegeneration, the ability of the PCC to maintain the error-monitoring function would become progressively impaired, accompanied by a reduction in Pe amplitude and progressive loss of awareness, as manifested by anosognosia from the prodromal AD stage.

Not least, anosognosia has been correlated with less depression and anxiety, but greater apathy^54,55, thus suggesting a potential common neurophysiological mechanism with partial overlap of neural circuits underlying anosognosia and these neuropsychiatric symptoms. For instance, emotional apathy, characterized by the inability to link affective and emotional signals to ongoing and upcoming behaviours, and to express empathy and interest, has been associated with dysfunction in the OFC and limbic territories of the basal ganglia (namely the ventral striatum)^56,57, negatively impacting error-monitoring that seems central in the mechanistic cascade leading to anosognosia. In fact, growing evidence points to an association between apathy in AD and hypometabolism and hypoperfusion in the ACC and OFC^58,59, as well as to a correlation between apathy and amyloid burden⁶⁰. Conversely, heightened awareness, instead of anosognosia, could contribute to the manifestation of depressive and anxious symptoms, with evidence showing an increase in Pe amplitude when subjects are over-concerned about errors^61,62. Remarkably, a recent study further revealed an association between more depressive symptoms with greater biomarkers of AD in individuals with only subjective cognitive decline, underscoring the relevance of assessing neuropsychiatric symptoms from the preclinical phase of this neurodegenerative disorder⁶³.

Main arguments supporting our perspective

The PCC and its cross-functional role underlying cognition and behaviour

Recent methodological advances in graph theory and neural dynamics have provided insights into the functioning of the human brain from a network perspective. Central to this framework are key nodes ensuring efficient neuronal signaling and communication^64,65. Among these nodes, the PCC, nestled within the medial brain surface, forms part of the structural core of the human connectome^66,67 and emerges as a crucial hub in the intricate network underlying cognition and behaviour^68,69. It is characterized by dense connectivity to widespread brain regions, making it the gateway of various functions. In fact, PCC’s cytoarchitectural subsections exhibit connections to key brain areas, including the visual and auditory cortices, superior and inferior parietal regions, lateral and orbital frontal structures, and the medial temporal cortex⁷⁰. The functional significance of the PCC is underscored by its remarkably high level of glucose metabolism and cerebral blood flow, which is about 40% higher than the average brain structures⁷¹. This elevated metabolic activity is supported by a dual blood supply from the posterior and middle cerebral arteries, a feature that, combined with its deep anatomical location, protects the PCC from local severe ischemic and traumatic insults. Nevertheless, its high centrality among brain networks makes it susceptible to dysfunction and disconnection^72–75, with critical evidence indicating the PCC as a particularly vulnerable hub to Alzheimer’s pathology^76,77.

Functionally, the PCC is at the core of the default mode network (DMN)⁷⁸, a dynamic brain network involved in self-referential processing^79,80. This network exhibits heightened metabolic activity during periods of rest and wakeful reflection compared to when engaged in externally oriented cognitive tasks^81,82. The PCC has been found to facilitate the integration of internal and external transmodal stimuli, orchestrating the seamless transition between rest and task-oriented states within the DMN^83–85. In fact, while implicated in self-referential processes typically associated with the DMN, the PCC is also involved in controlling the focus of attention by estimating the significance of sensory inputs and guiding attentional resources for adaptive responses to changing situational demands^37,86,87 through its interactions with other large-scale brain networks^85,88,89. Hence, the PCC operates beyond the confines of the DMN^{68,84,85,90–92}, and is intricately involved in performance monitoring, error awareness and cognitive control mechanisms essential for problem solving and developing learning strategies^39,93.The PCC has been found to play a crucial role in reinforcement-based learning paradigms, where actions are reinforced or punished based on their outcomes^39,94. Through its connections with the dopaminergic system and reward-related brain regions, such as the ventral striatum and OFC, the PCC modulates the encoding of reward signals and the evaluation of action-outcome contingencies, thereby guiding decision-making and adaptive behaviour in challenging environments^94–98. Its interactions with the retrosplenial cortex and medial temporal lobe, involved in spatial navigation and memory processes^70,99–101, support the PCC’s contribution to the construction of coherent spatial and temporal representations of the self and the environment⁵⁰. This intricate interplay between the PCC and limbic regions underlines its involvement in the generation of subjective experiences^102,103, which is crucial for conscious awareness, with direct evidence showing that transient disruption of PCC connectivity induces a loss of contact with the external world¹⁰⁴. Bearing all this in mind, recent work proposes considering the PCC as a key player in the unfolding of cognition, rather than constraining it to a specific brain function¹⁰⁵.

Taken together, its deep anatomical location, combined with a dense vascular supply and high metabolic rate, along with its multiple interactions with diverse brain networks and its close association with conscious awareness, positions the PCC as a pivotal brain region with a crucial cross-functional role underlying cognition and behaviour.

The PCC throughout the AD spectrum and its relationship to anosognosia

Consistent with its role in a broad range of cognitive and behavioural processes, the PCC has been implicated in several neurological and psychiatric disorders^86,106–111 and often associated with anosognosia. In psychiatric populations, anosognosia is usually referred to as poor insight¹¹². Notably, stronger connectivity between the PCC and the ventromedial prefrontal cortex has been associated with poorer clinical insight in young subjects at high risk of psychotic disorders¹¹³ and in schizophrenic patients¹¹⁴. This reflects impaired self-awareness, as revealed by standard questionnaires and self-reflection tasks, in the context of DMN dysregulation. Moreover, there is also ERP study evidence of compromised error-monitoring abilities in schizophrenic patients¹¹⁵.

In AD, substantial evidence points to an early disruption of the DMN^116–120, with prominent involvement of the PCC¹²¹. In particular, higher amyloid-β levels in the PCC were associated with lower functional connectivity in the DMN¹²². This is of utmost relevance, as the accumulation of amyloid-β in the brain appears to be the primary triggering event in the Alzheimer’s cascade, leading to early synaptic dysfunctions^123,124. Along this line, PCC synaptic function is affected from the very early onset of AD¹²⁵, with further ERP findings suggesting that synaptic failure in the error-monitoring system is the critical mechanism for the emergence of anosognosia in this neurodegenerative disorder³². Additional support to this view comes from research showing an association between anosognosia and both increased amyloid-β burden and reduced glucose metabolism in the PCC of elderly subjects at risk of AD⁴⁶, with amyloid-β burden predicting longitudinal changes in awareness of memory abilities in subjects who will progress to dementia⁵¹.

Disturbances of memory abilities are also affected by PCC integrity. There is evidence of a relationship between altered functional modulation in the DMN and episodic memory performance, as well as a significant age effect characterized by a reduced ability to deactivate the PCC hub during encoding, thereby suggesting a potential link between PCC activity and age-related changes in episodic memory processing¹²⁶. Recent research on neuromodulation has confirmed a causal effect between stimulation of the PCC during encoding and subsequent impairment of memory recall¹²⁷. Other pieces of evidence based on longitudinal studies have also shown an early relationship between increased amyloid-β burden in PCC and a decline in episodic memory in subjects who were amyloid-β negative at study entry^128,129.

Compared to the hippocampus, classically identified as a critical hub of AD, the PCC appears to play a more pivotal role in the progression of the disorder^130,131. This aligns with the view that neurodegenerative diseases target specific brain networks^{116,118,132–134}. In support to this network degeneration perspective, amyloid-β burden has been found to correlate with an accelerated decline in white matter integrity of the parahippocampal cingulum in healthy elderly subjects at risk of AD, with further evidence reinforcing the link between amyloid-β pathology and myelin damage^135,136. No less important, recent findings revealed upregulation of PCC genes in elders who remained cognitively unimpaired despite Alzheimer’s pathology^137,138, strongly suggesting a specific PCC genes profile as a potential mechanism for cognitive resilience in advanced ageing^138,139.

In sum, these research findings clearly support the PCC role in the modulation of cognition and behaviour in the ageing process. Notably, they show that early PCC impairment may have a major impact on AD progression, in a close link with anosognosia. Conversely, upregulation of PCC genes and associated synaptic pathways may counteract cognitive decline in elderly subjects, even in the presence of Alzheimer’s pathology.

The PCC and anosognosia in distinct neurological conditions

The PCC serves as a key node in understanding the intricacies of anosognosia beyond AD, especially in comparison with the stroke model and other neurological conditions, such as traumatic brain injury (TBI). In stroke patients, particularly those with AHP, unawareness has been traditionally associated with the failure of motor performance monitoring^140,141, typically following a lesion in the premotor cortex¹⁴². However, recent investigations have unveiled a convergence of neural disruptions, notably disconnection within the DMN, frontal latero-medial, fronto-striatal and fronto-parietal networks^21,143. These findings confirm the role of motor error monitoring and reveal the contribution of additional functional systems. These systems extend beyond motor performance to address the patients’ inability to update their self-referential beliefs about their condition¹⁴⁴, shift and direct attention towards salient stimuli¹⁴⁵, and manage reality monitoring failures¹⁴⁶. Hence, although not directly lesioned, the PCC and its functional networks become disconnected, contributing to anosognosia.

The disconnection of the PCC also plays a role in anosognosia for visual deficits, further confirming that, while spared by the stroke lesion, its indirect involvement is necessary for the emergence of unawareness of symptoms²². In TBI, anosognosia is often investigated through the lens of online and offline performance monitoring¹⁴⁷. TBI patients frequently exhibit impairments in accurately evaluating their performance across various cognitive tasks. Neuroimaging studies have revealed the involvement of the fronto-parietal control network and the cingulate cortex in mediating TBI-induced anosognosia¹⁴⁸. These regions play key roles in executive functions and attentional allocation, underscoring their relevance in the manifestation of anosognosia following TBI. In this complex neuroanatomical landscape, the PCC thus emerges as a potential integrative hub, facilitating communication between disparate neural systems that serve several cognitive components of self-awareness. The obvious drawback of such an important hub is that it can have a dramatic impact on the functional integrity of several brain networks when damaged by even a supposedly minor lesion.

A crucial disparity between the temporal dynamics of anosognosia in stroke and AD lies in their trajectories. While anosognosia in stroke often exhibits a transient nature¹⁴⁹, diminishing over time as patients recover from their deficits, it tends to worsen over the AD course. A lack of awareness for memory deficits in AD patients has been associated with critical synaptic failure in the error-monitoring system, in which the PCC plays a fundamental role. Altogether, this might suggest that synaptic changes in this specific brain region, resulting from either direct or indirect damage of distinct etiologies, could determine the evolution of anosognosia over time³². In AD, pervasive neurodegeneration undermines the brain’s capacity for error-based learning, limiting patients’ ability to adjust behaviour and compensate for their cognitive impairments. Conversely, early synaptic plasticity in stroke allows for the patients to engage in compensatory mechanisms^150,151 and adaptive strategies to mitigate the impact of their deficits¹⁵². This is in line with the view that anosognosia is error-prone; that is, it manifests in the presence of errors committed in the context of patients’ deficits³². Importantly, recent technological advances in promoting neuroplasticity hold great promise for enhancing recovery from brain damage of various etiologies¹⁵³.

Overall, these research findings indicate the involvement of monitoring and attentional networks in mediating anosognosia across diverse etiologies, further emphasizing the potential significance of PCC as a convergent node within the neural substrates underlying impaired self-awareness in distinct neurological conditions.

Anosognosia as a driving force in the clinical progression of AD: a disorder-network perspective centred on the posterior cingulate cortex

The foregoing arguments demonstrate the interconnected network nature of the brain and the early involvement of the PCC in the AD course. It thus seems very plausible to envisage Alzheimer’s pathophysiology from the perspective of a large-scale brain network^130,154,155, in which the PCC would play a pivotal role due to its extensive anatomo-functional connectivity (Fig. 2a). Specifically, we predict that primary damage to the PCC would be critical for the emergence of anosognosia from the early stages of AD, which would, in turn, determine the clinical progression of the disorder. More precisely, we posit that an inaugural, early AD-related event in the PCC, namely amyloid-β deposits (Fig. 2b), would be at the origin of anosognosia in AD patients, through a metabolic deficit in the PCC (Fig. 2c) and a synaptic failure in the error-monitoring system, resulting in their inability to consciously perceive errors made in the context of cognitive deficits (Fig. 2d).

Fig. 2. The neurofunctional cascade of anosognosia in Alzheimer’s disease.

a Green arrows represent the medial functional connectivity of the PCC in the healthy brain. The anatomical borders of the PCC are identified as in the Harvard-Oxford Atlas¹⁷⁹. b Representation of amyloid-PET imaging positivity in the PCC of an AD brain. The colour bar indicates amyloid-PET tracer uptake, ranging from low (blue) to high (red). Amyloid-β deposits are indicated by black dots. c Representation of hypometabolism pattern in the PCC of an AD brain, as measured by FDG PET. The colour bar represents metabolic activity, measured as FDG uptake, from low (blue) to high (red). d Schematic representation of a decrease in Pe amplitude recorded in the PCC of an AD brain. Downward-pointing black arrow indicates Pe amplitude decrease. Red arrows show the medial view of the impairment of functional systems connected to the impaired error-monitoring system within the PCC hub. Aβ amyloid-β. AD Alzheimer’s disease, ERN error-related negativity. FDG F-fluorodeoxyglucose, PCC posterior cingulate cortex, Pe error positivity.

Anosognosia would then prevent patients from learning from their errors and adopting strategies to mitigate their deficits, further disrupting the access to episodic memory retrieval. This would contribute decisively to cognitive decline and drive the clinical progression of Alzheimer’s patients (Fig. 3). Hence, at the very early stages of the disorder, the worsening of AD symptoms would critically depend on whether patients consciously perceive their cognitive deficits, with the occurrence of anosognosia determining the patients’ inability to implement strategies to tackle their difficulties, thus feeding the vicious circle of cognitive decline.

Fig. 3. The vicious circle of cognitive decline in Alzheimer’s disease.

A failure in the error-monitoring system, secondary to PCC damage, is at the root of patients’ anosognosia. This lack of awareness of errors committed in the context of their deficits prevents learning from those errors. This, in turn, prevents behavioural adjustments, leading to the inability to implement strategies to cope with the deficits, with consequent worsening of symptoms. This process perpetuates a vicious circle of cognitive decline, driving the clinical progression of Alzheimer’s disease.

Discussion and future outlook

Alzheimer’s disease constitutes a global healthcare challenge, primarily because age is its main risk factor and people are living longer¹⁵⁶. This issue is exacerbated by the lack of effective therapies to modify the underlying neurodegenerative process and reverse the dramatic cognitive decline that characterizes the disease¹⁵⁷. In fact, even though the first disease-modifying treatments are sowing hope, innovative and complementary interventions are urgently needed to curb this public health crisis¹⁵⁸.

Known for more than a century, the two cardinal pathologies of AD are sine qua non for confirming the diagnosis of this neurodegenerative disorder^159,160. Nevertheless, the relationship between the level of Alzheimer’s pathology and cognitive decline is more complex than previously thought^161,162. For instance, some individuals retain their cognitive abilities despite the presence of high levels of Alzheimer’s pathology in their brains^163,164. Thus, it is not yet clear whether all subjects with preclinical AD will manifest symptoms conducive to AD diagnosis. Indeed, there are still important gaps in our knowledge of the neural mechanisms contributing to cognitive decline over the AD continuum. Memory loss that typically characterizes the prodromal stage of the disorder mostly affects episodic memory¹⁶⁵, which is part of the declarative long-term memory system that requires a conscious recall process^166,167. Crucial evidence links episodic memory deficits, revealed by tests demanding an effortful retrieval monitoring, to anosognosia in AD patients⁴⁹.

Here, we propose that anosognosia represents a major determinant in the clinical trajectory of AD, exerting a causal role in symptom progression and patient outcomes. In particular, our disorder-network perspective centres on the PCC, positing anosognosia as a primary driver of cognitive decline in AD, with the PCC as its pivotal node following early amyloid-β deposits in this specific brain region. We advance the novel view that anosognosia, resulting from a failure in the error-monitoring mechanisms supervised by the PCC, impairs patients’ awareness of their own condition. This lack of awareness prevents them from acknowledging and learning from their errors, adjusting their behaviour and retrieving episodic memories, thereby triggering and perpetuating a vicious circle of cognitive decline. This paradigm shift underscores the critical role of impaired self-awareness in exacerbating cognitive deficits and shaping the course of the disease, giving anosognosia a unique role in the AD continuum. The originality of our perspective is to consider anosognosia not as a mere risk factor or even a predictor of AD progression but rather as the main driver of this progression within a disorder-network centred on the PCC—a brain region, with distinctive biological and anatomo-functional features, playing a prominent role in cognition and behaviour, which has also been shown vulnerable to AD, exhibiting amyloid-β deposits^120,168 and hypometabolism¹⁶⁹ from the earliest stages of the disorder⁷⁶.

The observation that some patients remain aware of their deficits even at an advanced stage of AD brings to the fore the variability of clinical presentations, in agreement with the view that the changes in awareness occurring along the AD continuum depend on a broad brain network in which the PCC plays a primordial and integrative role, but is not the sole contributor, as illustrated by the dual-path hypothesis for the emergence of anosognosia³². In fact, the motivation required to engage in self-regulation and the emotional trigger that marks errors as significant are aspects of performance monitoring that are often disregarded. It is quite possible that other regions, such as those involved in sustained attention, reward and emotional regulation, provide compensatory mechanisms that preserve awareness in certain individuals. The persistence of awareness in some cases could also reflect individual differences in neuronal reserve, synaptic plasticity or structural connectivity that help maintain self-referential processing despite significant pathology.

Of note, contrary to the commonly held idea that anosognosia might shield patients from the psychological distress associated with their deficits, our perspective proposes a different understanding of its impact. Unaware of their deficits, patients are no longer able to take care of themselves and engage in meaningful strategies to cope with their difficulties, which might aggravate their condition.

The intricate interplay between anosognosia, error-monitoring-induced learning deficits and cognitive decline underscores the need for targeted interventions aimed at improving awareness and mitigating disease progression¹⁷⁰. Growing evidence on the neural mechanisms underlying anosognosia in different neurological conditions, including AD, stroke or TBI, suggests a universal role of PCC and related error-monitoring mechanisms in the emergence of this phenomenon. These shared anatomical and neural mechanisms of anosognosia could translate into novel strategies for promoting self-awareness in AD by tapping, for example, into successful training programs implemented for AHP, such as third-person perspective taking and video recording and evaluation of one’s own behavioural performance. Additionally, the neuromodulation of the DMN and the salience network targeting the PCC has shown encouraging results in the treatment of psychopathologies¹⁷¹, pointing to the use of neuromodulation techniques, such as transcranial magnetic stimulation and deep brain stimulation, or even neurofeedback¹⁷², as a potential therapeutic approach to modulate functional brain networks that regulate self-awareness from the preclinical stages of AD. Furthermore, there is causal evidence implicating the PCC in episodic memory processes¹²⁷, making this brain region a relevant target for neuromodulation strategies aimed at improving memory abilities in subjects at risk of AD¹⁷³. Noteworthy, neuromodulation underpins the flexibility of neural circuitry by targeting synaptic dynamics, thereby shaping electrophysiological activity¹⁷⁴. Synaptic failure is an early event in the pathogenesis of AD which, in most early stages, engages neural compensatory mechanisms, suggesting an initial reversible stage of downregulation preceding a stage of decompensated and irreversible function^123,124,175. Strategies aimed at regulating synaptic function from the earliest, even preclinical, stages of AD could therefore prove extremely valuable in the ongoing efforts to tackle cognitive decline in the elderly^176,177.

Last, it is important to distinguish between AD and dementia. Dementia is only the late stage of a disease that takes decades to develop in the human brain. Thus, the identification of anosognosia from the very early stages of AD constitutes a critical window of opportunity for intervention over its long continuum¹⁷⁸. Electrophysiological signals related to error awareness changes could serve as a diagnostic or predictive biomarker along the AD continuum, representing a promising target for neuromodulation in subjects at risk of developing this dramatic disorder. By conceiving meaningful approaches targeting patients’ awareness mechanisms, we could help them implement strategies to counter their deficits, enhance conscious access to episodic memory retrieval, adjust their behaviour, and maintain their connectedness with the external world. This could potentially delay cognitive decline and enable patients to avoid risky situations. As a result, patients could also become more receptive to therapeutic interventions and more willing to engage in social interactions, improving their quality of life, with positive repercussions for caregivers and healthcare systems too.

In conclusion, while we have gathered significant evidence from many fields of cognitive neuroscience to support our perspective, further studies are needed to investigate the causal role of anosognosia (and its roots in the PCC) as the main driver of the clinical decline over the AD course. Validating this novel perspective opens up innovative avenues for a deeper understanding of the pathophysiology of AD and the development of complementary therapeutic interventions. In this respect, exploring whether improved self-awareness can activate neural mechanisms of cognitive resilience, preventing or at least delaying the transition from the preclinical to the symptomatic AD phase, holds promise for improving Alzheimer’s care.

Author contributions

K.A. established the theoretical foundations of the perspective. K.A. and V.P. contributed to the conceptualization of the manuscript, writing of the first draft and revisions. Both authors have read and approved the final version of the manuscript.

Peer review

Peer review information

Communications Biology thanks Patrizia Vannini and Tomoyuki Nagata for their contribution to the peer review of this work. Primary handling editor: Benjamin Bessieres. [A peer review file is available.]

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s42003-024-07076-7.