Mid-life anxiety or depression increase the risk of developing dementia in people with mild cognitive impairment, and these psychiatric symptoms worsen as Alzheimer’s disease progresses. Now, researchers led by Yonas Geda, Barrow Neurological Institute, Phoenix, report that psychiatric symptoms also correlate with AD biomarkers in cognitively normal older adults. In the February 9 Alzheimer’s & Dementia, they reported that, even though they were cognitively normal, people who were anxious, depressed, or apathetic had less Aβ42 in their cerebrospinal fluid and a higher tau/Aβ42 ratio than people without those complaints.

  • People with high tau/Aβ42 ratios were threefold likelier to have anxiety.
  • They were also twice as likely to have apathy and sleep disturbances.
  • Could treating neuropsychiatric disorders prevent or slow cognitive decline?

Though not the first to report such a relationship, the paper adds weight to it by studying multiple psychiatric symptoms in a relatively large cohort. The study begs the question of which comes first: the symptoms or the biomarkers? A hint comes from a paper in the February 16 JAMA Psychiatry. Researchers led by Leah Richmond-Rakerd, University of Michigan, Ann Arbor, reported that among 1.7 million New Zealanders, those with early life mental disorders were at higher risk for dementia and they were more likely to have it at a younger age.

“Neuropsychiatric symptoms are associated with AD core pathology, suggesting that this pathology is at least partly driving their occurrence,” Constantine Lyketsos, Johns Hopkins University, Baltimore, wrote to Alzforum. Lee Hyer, Mercer School of Medicine, Savannah, Georgia, thinks that the appearance of neuropsychiatric disorders may be another way to identify early neurodegeneration, but which symptoms and how to best identify them needs further evaluation, he said (full comment below).

Behavioral symptoms are well-known to accompany mild cognitive impairment and AD (Jan 2015 news). Anxiety, depression, agitation, and apathy also have been linked to abnormal AD biomarkers, albeit not always at the same time and with varying degrees of confidence. Of three studies on MCI participants in the Alzheimer’s Disease Neuroimaging Initiative, one tied low CSF Aβ42 and high total tau to anxiety, agitation, and irritability, another found abnormal biomarkers only in those with anxiety and apathy, and the third linked only low Aβ42 to chronic depression (Ramakers et al., 2012; Banning et al., 2020; Gonzales et al., 2018). In a recent analysis of 24,500 ADNI and National Alzheimer’s Coordinating Center participants ranging from cognitively normal to AD dementia, those with low CSF Aβ42 and high t-tau were likelier to develop depression or apathy over five years (Banning et al., 2021).

How do fluid AD biomarkers correlate with neuropsychiatric symptoms in cognitively normal older adults? To find out, co-first authors Janina Krell-Roesch and Martin Rakusa, who worked with Geda when he was at the Mayo Clinic, Rochester, Minnesota, turned to data from 699 cognitively normal older adults and 85 who had MCI from the Mayo Clinic Study of Aging (MCSA). The average age was 73, and 57 percent were men.

The scientists compared CSF Aβ42, total tau, and phosphotau-181 to scores on the self-reported Beck Anxiety and Depression Inventories (BAI and BDI) and on the Neuropsychiatric Inventory Questionnaire. The NPI-Q, taken by a caregiver, asks about the presence or absence of 12 symptoms, including anxiety, depression, apathy, agitation, and irritability. It also asks about changes in nighttime behavior, such as sleeplessness, sleepwalking, or having nightmares.

In correlating biomarkers and symptoms, the authors adjusted for age, sex, years of education, and APOE4 genotype. As expected, participants with MCI had less Aβ42 in their CSF and more t-tau and p-tau181, and were more anxious or depressed than their cognitively normal counterparts. In the cohort as a whole, people classified as clinically anxious or depressed, namely those with BAI scores over 10 or BDI scores over 13, respectfully, also had abnormal AD markers—they were twice as likely to have high t-tau/Aβ42 and p-tau181/Aβ42 ratios.

But what about cognitively normal people? Among them, the more anxious or depressed a person was, the lower his or her Aβ42 and higher the t-tau/Aβ42 and p-tau181/Aβ42 ratios. People with high tau/Aβ42 ratios were three times as likely to be categorized as anxious by the NPI-Q, 2.5-fold more likely to have sleep problems, and twice as likely to be apathetic.

The scientists had previously tied high anxiety and depression to positive amyloid PET scans and lower brain glucose metabolism in cognitively normal participants from this same cohort (Krell-Roesch et al., 2018; Krell-Roesch et al., 2021; Krell-Roesch et al., 2016).

[Courtesy of Janina Krell-Roesch and Yonas Geda.]

All told, the data suggest a link between neuropsychiatric symptoms and AD biomarkers in adulthood, but which comes first? Geda and colleagues proposed four scenarios (Geda et al., 2013). One: Psychiatric disorders may damage the brain through unknown mechanisms, increasing risk of AD. Two: Neuropsychiatric symptoms and AD develop independently from shared risk factors for both, such as genetic variants (Lutz et al., 2020). Three: Reverse causality, i.e. mild cognitive trouble early in AD, makes people anxious and depressed, which accelerates dementia. Four: Neuropsychiatric problems synergize with AD risk factors to speed dementia onset (see image below).

Longitudinal studies could sort out which of these scenarios is true. One analysis of 1,440 cognitively normal MCSA participants followed for 5.5 years found that being both anxious and amyloid-positive via PET compounded a person’s risk of MCI (Pink et al., 2021). Similarly, among 104 cognitively normal and 53 people with MCI in the Swedish BioFINDER study, those with anxiety or apathy had a higher plaque load, more cerebral atrophy, and declined faster (Johansson et al., 2020). In a more recent study of 350 participants in BioFINDER, amyloid-positive people were more likely to be anxious and apathetic, and greater apathy sped cognitive decline over eight years (Johansson et al., 2022). The reciprocal did not hold, however. “The effect of CSF Aβ42 status on longitudinal change in neuropsychiatric symptoms was mainly independent of cognitive decline,” wrote Oskar Hansson and Maurits Johansson of Lund University, Sweden, who led the study (full comment below).

Separately, the large, longitudinal study in New Zealand suggests that psychiatric symptoms increase risk for AD. Examining the medical records of 1.7 million adults from 1988 to 2018, Richmond-Rakerd and colleagues concluded that mental disorders, even in early life, often precede and accelerate dementia. People with anxiety or depression were three times as likely to be diagnosed with dementia, while those with schizophrenia or a related disorder had a sixfold higher risk (see image below). Having any mental condition brought on dementia 5.6 years sooner than not having one.  

Mental Health and Dementia. People with mental illness were three to six times more likely to develop dementia (top), including AD (bottom). [Courtesy of Richmond-Rakerd et al., JAMA Psychiatry, 2022. ©2022 American Medical Association. All rights reserved.]

“People with neuropsychiatric symptoms are at high risk of progressive cognitive decline, so there is an opportunity to delay decline by targeting these symptoms early,” Lyketsos wrote. Both he and Wes Ashford at Stanford University, California, pointed to the degeneration of neurons rich in serotonin and norepinephrine and subsequent neurotransmitter signaling breakdown as operative processes in depression and AD. “Both diseases are likely related to similar underlying damage of these neurons,” Ashford believes (full comment below).

Could treating neuropsychiatric symptoms improve cognition, or even stave off AD? On the former, the data is mixed. The norepinephrine reuptake inhibitor atomoxetine increased norepinephrine in the brain, enhanced glucose metabolism, and lowered CSF t-tau and p-tau181 levels in a small trial of 39 people with MCI (Levey et al., 2021). Neither CSF Aβ42 nor cognition changed after 29 weeks on this ADHD drug.

In two mild to moderate AD trials, 12 weeks of the norepinephrine-dopamine reuptake inhibitor bupropion improved neither apathy nor cognition. In contrast, 12 weeks of the selective serotonin reuptake inhibitor sertraline reduced depression and improved cognition in people with moderate AD (Maier et al., 2020; Mokhber et al., 2014). An ongoing retrospective study is exploring if people with MCI who have been taking SSRIs for at least six months progress to dementia more slowly than those who haven’t.—Chelsea Weidman Burke


  1. The Krell-Roesch et al. study presents a rich set of results on the value of biomarkers and behavioral indices on depression and anxiety in a group of normal and MCI subjects. The study contributes to this body of research by providing preliminary evidence that lower CSF Aβ42, and higher t-tau/Aβ42 and p-tau/Aβ42, are associated with depression and anxiety as measured in different ways by scales. Increased odds of presence of some behaviors, especially apathy and altered nighttime behavior, were also noteworthy.

    The authors used a large sample of community-dwelling older adults free of dementia, detailed assessment of neuropsychiatric symptoms using both self-reported BDI-II and BAI, as well as informant-observed NPI-Q. The study did allude to mild behavioral impairment (MBI) as being associated with AD biomarker abnormalities.  

    The neuropsych battery used here is solid and representative of the cognitive domains. The behaviors were evaluated by the NPI and via an informant. I believe the MBI may be a better marker that the NPI, as this measure is discrete. Clearly one can consider depression and anxiety measured by scales as behaviors. In this case, the study “validated” the need for behaviors in the assessment. 

    Now I believe the question becomes what is the best set of behaviors to be assessed and how to assess them? The MBI has 10 or so studies on the issue of neurodegeneration and is a more promising index to assess early signs/symptoms of this problem. The NPI assesses more severe markers of neurodegeneration. It is examiner-based, which can be an advantage or a problem, subject to the strengths/weaknesses of an interviewer versus a self-report measure. Additionally, the MBI has more than 30 questions assessing five more relevant behaviors in the earlier stages of a decline process. It can also be rated by a significant other.

  2. This paper makes a very important contribution to the literature by investigating the cross-sectional associations between core AD-related biomarkers in CSF and neuropsychiatric symptoms (NPS) in a large cohort of non-demented people.

    The research field of NPS in AD struggles with diverse methodologies and conflicting findings. The straightforward methodology in this paper is therefore much welcomed and open for easier comparisons with other studies. Their key findings, including associations between Aβ pathology and certain NPS, such as anxiety and apathy, are relatively consistent with the overall literature. For example, we have previously demonstrated that anxiety and apathy are associated with Aβ deposition by PET imaging in non-demented individuals (Johansson et al, 2020). 

    However, the literature on depression is less conclusive. This probably reflects the polythetic concept of depression, which is captured somewhat differently between various rating scales. This is clearly illustrated by the opposing findings in this study when using the BDI-II or the NPI subdomain “dysphoria.” Therefore, it is a strength of this study that it studied associations with “clinical depression” obtained by dichotomization of the BDI-II according to a clinical cut-off. This probably increases the likelihood of having addressed “depression” and not NPS with overlapping symptomatology such as apathy or anxiety. In line with the use of the continuous BDI-II total rating score, clinical depression was related to lower levels of CSF Aβ42, as well as higher levels of the ratios of P-tau/Aβ42, and T-tau/Aβ42, respectively.

    The interplay between NPS, cognition, and AD pathologies is complex. As the authors comment, the causal direction of the demonstrated associations cannot be displayed due to the study’s cross-sectional nature. Interestingly, the stratified analyses according to cognitive status show more robust associations between NPS and CSF Aβ42 among the MCI than the CU subjects, which indicates that cognitive dysfunction could play a role in NPS evolution.

    Accordingly, studies exploring the effect of AD pathologies should control for cognitive dysfunction in order to reduce confounding.

    In a recent paper, we demonstrated that the effect of CSF Aβ42 status on longitudinal change in NPS was mainly independent of cognitive decline (Johansson et al, 2022). However, additional longitudinal studies in this area are needed.


    . Apathy and anxiety are early markers of Alzheimer's disease. Neurobiol Aging. 2020 Jan;85:74-82. Epub 2019 Oct 19 PubMed.

    . Development of apathy, anxiety, and depression in cognitively unimpaired older adults: effects of Alzheimer’s disease pathology and cognitive decline. Biological Psychiatry, 31 January 2022

  3. There is a well-established relationship between Alzheimer’s disease and depression (Ashford, 2018). 

    Below are the “risk factors” for Alzheimer’s disease, citations below; depression is listed: APOE-e4 genotype - 1 allele x 4; 2 alleles x 16

    • Family history of dementia - 3.5 (2.6 - 4.6)
    • Family history - Downs 2.7 (1.2 - 5.7)
    • Family history - Parkinson’s 2.4 (1.0 - 5.8)
    • Obese, large abdomen - 3.6
    • Maternal age > 40 years - 1.7 (1.0 - 2.9)
    • Head trauma (with LOC) - 1.8 (1.3 - 2.7)
    • History of depression - 1.8 (1.3 - 2.7)
    • History of hypothyroidism- 2.3 (1.0 - 5.4)
    • History of severe headache - 0.7 (0.5 - 1.0)
    • History of “statin” use - 0.3
    • NSAID use - 0.2 (0.05 – 0.83)
    • Use of NSAIDs, ASA, H2-blockers - 0.09

    There is even some thought that there is a genetic link between depression and AD (Lutz et al., 2020). The focus on genetic factors is interesting but does not tell the whole story.

    Two changes characteristic of very early Alzheimer’s disease are, first, loss of norepinephrine neurons in the locus coeruleus of the dorsal pons in the brainstem and, second, loss of serotonin neurons in the midbrain raphe system. Both norepinephrine and serotonin are thought to be related to depression since all of the anti-depressant medications work through one or both of these neurotransmitter systems (Ashford, 2018). And the norepinephrine and serotonin neurons are also involved in memory mechanisms (Ashford and Larvik, 1985). 

    The problem of the Alzheimer's and depression relationship in the elderly, not young individuals, is that both are likely related to similar underlying damage of the norepinephrine and serotonin neurons. The damage in the brainstem is likely the initial factor, which leads to both depression and Alzheimer's disease. Norepinephrine and serotonin synapses in the cortex, releasing their neurotransmitters in response to operant conditioning and classical conditioning signals respectively, activate the α-secretase (ADAM-10) to cleave the amyloid-protein precursor and stimulate the formation of new synapses. When an activated neuron does not get activation of the α-secretase, the default β-secretase pathway leads to removal of the local synapse, which is destroyed by the γ-secretase product Aβ, while the synaptic spine is withdrawn by the complement component by activating phosphorylation of the microtubule-associated protein-TAU. This careful balance leads to a turnover of synapses in the brain at a rate of about 5 trillion synapses (made and destroyed) per day.

    Depression has been related to underactivity of norephinephrine and/or sertonin neurons. In depression, when it is an early symptom of Alzheimer's disease, the damaged projections of these neurons to the frontal lobes fail to activate the α-secretase to cleave the amyloid pre-protein, leading to excess of the default β-secretase and formation of excess amyloid, which is seen as deposits in early phases of Alzheimer's disease predominantly in the frontal lobes, though without dementia, but unrecognized cognitive changes likely do occur. However, the actual deposition of amyloid is probably not relevant, with the depression related to failure of frontal lobe processing of new information related to the decrease of α-secretase action and the excess loss of synapses due to the actions of the default β pathway.

    Dementia occurs later, with more severe impairment of the norepinephrine and serotonin pathways. The failure of norepinephrine and serotonin activation in the regions of cortex with heavy memory activity and neuroplasticity, the hippocampus and amygdala and posterior temporal and inferior parietal cortices, leads to lack of processing of incoming information and memory dysfunction. These regions appear to be affected at a later age and predominantly show the signs of excess Tau phosphorylation, and these regions manifest less amyloid deposition, likely due to their adaptation for heavy amounts of neuroplasticity.

    In any case, the amyloid protein precursor appears to be important, as overexpression of the gene encoding this protein is responsible for the dementia of Down’s syndrome, and the predisposition to amyloid deposition is closely related to APOE genotype, as evidenced by the age of onset of amyloid deposition and decades later by the age of onset of dementia. But neither of these points proves that amyloid deposition is related either to depression or dementia. In fact, while CSF amyloid goes down in relationship to APOE-e4 genotype, it is not related to dementia severity. Alternatively, CSF TAU increases in relationship to dementia severity but is not related to APOE genotype. It is more likely the amyloid deposition is serving as a scar caused by the underlying neurochemical process, which causes the underproduction of synapses and their over-destruction.

    This view suggests that there is no role for amyloid removal as a treatment for Alzheimer's disease or its symptoms. The important questions are why the norepinephrine and serotonin neurons deteriorate, and how to alleviate that degeneration. The current best answers are healthy behaviors, though precision application according to genotype needs extensive study.


    . The Dichotomy of Alzheimer's Disease Pathology: Amyloid-β and Tau. J Alzheimers Dis. 2019;68(1):77-83. PubMed.

    . Shared genetic etiology underlying Alzheimer's disease and major depressive disorder. Transl Psychiatry. 2020 Mar 9;10(1):88. PubMed.

    . Alzheimer's disease: does neuron plasticity predispose to axonal neurofibrillary degeneration?. N Engl J Med. 1985 Aug 8;313(6):388-9. PubMed.

    . Frequency, distribution, and risk factors for Alzheimer's disease. Nurs Clin North Am. 1994 Mar;29(1):101-11. PubMed.

    . Nonsteroidal antiinflammatory drugs and the risk of Alzheimer's disease. N Engl J Med. 2001 Nov 22;345(21):1515-21. PubMed.

    . Risk of Alzheimer disease with the epsilon4 allele for apolipoprotein E in a population-based study of men aged 62-73 years. Alzheimer Dis Assoc Disord. 1998 Mar;12(1):40-4. PubMed.

    . Decreased prevalence of Alzheimer disease associated with 3-hydroxy-3-methyglutaryl coenzyme A reductase inhibitors. Arch Neurol. 2000 Oct;57(10):1439-43. PubMed.

Make a Comment

To make a comment you must login or register.


News Citations

  1. Behavioral Symptoms Hit Early in Alzheimer’s

Therapeutics Citations

  1. Atomoxetine

Paper Citations

  1. . Anxiety is related to Alzheimer cerebrospinal fluid markers in subjects with mild cognitive impairment. Psychol Med. 2012 Sep 7;:1-10. PubMed.
  2. . The Association Between Biomarkers and Neuropsychiatric Symptoms Across the Alzheimer's Disease Spectrum. Am J Geriatr Psychiatry. 2020 Jul;28(7):735-744. Epub 2020 Feb 20 PubMed.
  3. . Chronic depressive symptomatology and CSF amyloid beta and tau levels in mild cognitive impairment. Int J Geriatr Psychiatry. 2018 Oct;33(10):1305-1311. Epub 2018 Jun 28 PubMed.
  4. . Alzheimer's disease biomarkers as predictors of trajectories of depression and apathy in cognitively normal individuals, mild cognitive impairment, and Alzheimer's disease dementia. Int J Geriatr Psychiatry. 2021 Jan;36(1):224-234. Epub 2020 Sep 11 PubMed.
  5. . Depressive and anxiety symptoms and cortical amyloid deposition among cognitively normal elderly persons: the Mayo Clinic Study of Aging. Int Psychogeriatr. 2018 Feb;30(2):245-251. Epub 2017 Dec 4 PubMed.
  6. . Association of Cortical and Subcortical β-Amyloid With Standardized Measures of Depressive and Anxiety Symptoms in Adults Without Dementia. J Neuropsychiatry Clin Neurosci. 2021;33(1):64-71. Epub 2020 Oct 22 PubMed.
  7. . FDG-PET and Neuropsychiatric Symptoms among Cognitively Normal Elderly Persons: The Mayo Clinic Study of Aging. J Alzheimers Dis. 2016 Jul 14;53(4):1609-16. PubMed.
  8. . Neuropsychiatric symptoms in Alzheimer's disease: Past progress and anticipation of the future. Alzheimers Dement. 2013 Apr 3; PubMed.
  9. . Shared genetic etiology underlying Alzheimer's disease and major depressive disorder. Transl Psychiatry. 2020 Mar 9;10(1):88. PubMed.
  10. . A longitudinal investigation of Aβ, anxiety, depression, and mild cognitive impairment. Alzheimers Dement. 2021 Dec 8; PubMed.
  11. . Apathy and anxiety are early markers of Alzheimer's disease. Neurobiol Aging. 2020 Jan;85:74-82. Epub 2019 Oct 19 PubMed.
  12. . Development of apathy, anxiety, and depression in cognitively unimpaired older adults: effects of Alzheimer’s disease pathology and cognitive decline. Biological Psychiatry, 31 January 2022
  13. . A phase II study repurposing atomoxetine for neuroprotection in mild cognitive impairment. Brain. 2021 Dec 17; PubMed.
  14. . Bupropion for the Treatment of Apathy in Alzheimer Disease: A Randomized Clinical Trial. JAMA Netw Open. 2020 May 1;3(5):e206027. PubMed.
  15. . Comparison of sertraline, venlafaxine and desipramine effects on depression, cognition and the daily living activities in Alzheimer patients. Pharmacopsychiatry. 2014 Jul;47(4-5):131-40. Epub 2014 Jun 23 PubMed.

External Citations

  1. study

Further Reading


  1. . Neuropsychiatric Symptoms and Neuroimaging Biomarkers in Alzheimer Disease: "Which is the Cart and Which is the Horse?". Am J Geriatr Psychiatry. 2017 Jul;25(7):694-696. Epub 2017 Mar 16 PubMed.
  2. . Cerebrospinal fluid biomarkers for neuropsychological symptoms in early stage of late-onset Alzheimer's disease. Int J Neurosci. 2014 Oct 29; PubMed.

Primary Papers

  1. . Association between CSF biomarkers of Alzheimer's disease and neuropsychiatric symptoms: Mayo Clinic Study of Aging. Alzheimers Dement. 2022 Feb 9; PubMed.
  2. . Longitudinal Associations of Mental Disorders With Dementia: 30-Year Analysis of 1.7 Million New Zealand Citizens. JAMA Psychiatry. 2022 Apr 1;79(4):333-340. PubMed.