![[Adobe Stock]](https://sp-ao.shortpixel.ai/client/to_webp,q_glossy,ret_img,w_750,h_375/https://www.psypost.org/wp-content/uploads/2026/11/brain-1-750x375.jpg)
In older adults without memory problems, the buildup of Alzheimer’s-related tau protein in the brain tends to precede and predict a rise in mild depressive symptoms, with little evidence that mood changes drive protein accumulation. The findings, published in the Journal of Neuroscience, provide evidence that subtle emotional shifts in late life may be an early consequence of Alzheimer’s-related brain changes rather than an independent trigger for disease progression. This pattern was statistically significant only in cognitively healthy individuals, suggesting that early temporal windows offer the best opportunity to observe how brain pathology affects mood.
Alzheimer’s disease (AD) is a progressive brain disorder that slowly destroys memory and thinking skills. It is biologically defined by the accumulation of two abnormal proteins: amyloid-beta, which forms sticky plaques outside brain cells, and tau, a protein that normally helps stabilize the internal skeleton of neurons but can misfold and form toxic tangles inside them. Long before a person shows noticeable memory loss, they often experience subtle changes in mood, such as apathy, anxiety, or mild depression.
Even mild depressive symptoms that fall short of a clinical diagnosis have been linked to a higher risk of developing dementia later on. Because tau accumulation and worsening mood often show up around the same time, it has remained an open question which process leads the other. One possibility is that early tau tangles damage deep brain regions that produce mood-regulating chemicals like serotonin, causing depressive symptoms to emerge as an early sign of disease.
Another possibility is that depression itself alters brain wiring or stress pathways in ways that make neurons more vulnerable to tau spread. Earlier imaging work linked the two in healthy aging. For example, a 2017 pilot study of cognitively normal older adults found that people with more tau tangles in memory-related brain areas also tended to report more depressive symptoms at a single point in time.
Don’t miss the next big psychology finding
Every morning, PsyPost sends a short briefing on the most interesting new peer-reviewed research in psychology and neuroscience.
Later longitudinal research tracked how these measures move together over several years. In a 2025 study by two of the same authors, Teodora Z. Markova and Anne S. Berry, analyzing shared data from the Alzheimer’s Disease Neuroimaging Initiative (ADNI), rising depressive symptoms in cognitively unimpaired older adults were linked to tau buildup in the amygdala, a brain hub that processes emotion. Still, standard correlation studies could not establish whether tau buildup was driving the mood changes or if worsening mood was accelerating tau accumulation.
“Most of the work thus far in the field was correlational, just showing associations between more tau pathology and worsening depressive symptoms,” study lead author Teodora Z. Markova, who conducted the research as a doctoral student in the Neurochemistry and Cognition Lab at Brandeis University and is now a postdoctoral researcher at the University of California, San Francisco Memory and Aging Center, told PsyPost.
“I wanted to leverage existing large-scale datasets like ADNI to really try and understand the dynamics between the two variables. While it’s hard to establish any ‘causal’ relationships with this kind of data, I wanted to use some creative statistical approaches to better understand which variable changes first, and get a better sense of whether tau accumulates before mood worsens (tau ‘causing’ disruptions in mood) or vice versa.”
To test the direction of this relationship, Markova, Berry at Brandeis University, and co-author Margo B. Heston at the University of California, San Francisco applied three statistical modeling approaches to multi-year brain scan and mood data. The authors originally hypothesized a two-way feedback loop in which tau and depressive symptoms would reinforce each other over time, with tau having the larger effect.
They examined records from 390 older adults (197 women and 193 men) enrolled in the Alzheimer’s Disease Neuroimaging Initiative, a large multi-site research project tracking brain aging. At the start of the tracking period, 226 participants had normal cognition, while 164 had either mild cognitive impairment (MCI), a stage of noticeable memory decline that does not yet interfere with daily independence, or Alzheimer’s dementia.
None of the participants had clinical depression when they entered the study, scoring 6 or lower on the 15-point Geriatric Depression Scale, a standard yes-or-no questionnaire that asks older adults about their mood and life satisfaction. Over an average follow-up period of roughly three years, each participant underwent at least two positron emission tomography (PET) brain scans to measure tau protein, paired with repeated depression questionnaires and memory tests. During a tau PET scan, a person receives a small injection of a radioactive tracer that binds to tau tangles in the brain, allowing a scanner to measure how much protein has accumulated in different areas.
Because tau spreads through the brain in a predictable sequence of stages, the researchers customized their measurements for each participant by tracking the most advanced brain region that had already crossed the threshold for tau positivity at their first scan. Most cognitively normal participants (194 of 226) had not crossed the positivity threshold anywhere at their first scan, so the researchers tracked tau in the earliest-affected region for them. They also took into account whether participants had amyloid-beta plaques on separate PET scans and whether they carried the APOE ε4 gene variant, the strongest known genetic risk factor for late-onset Alzheimer’s disease. The full statistical outputs were posted publicly on the Open Science Framework.
As expected, participants in the cognitively impaired group started out with more tau in their brains and more depressive symptoms than the cognitively healthy group, and both measures increased more rapidly over time in the impaired group. On the 15-point depression scale, the cognitively normal group started with an average score of 0.8, compared to 2.1 in the cognitively impaired group. Having amyloid plaques or carrying the APOE ε4 gene variant was linked to higher starting tau levels, particularly among those who already had cognitive impairment, though neither factor was significantly linked to differences in starting depression scores or how they changed over time.
When the researchers tested how rates of change in tau and mood interacted over time, a pattern pointing in one direction emerged in the cognitively normal group, running partly counter to their expectation of a two-way loop. Cognitively healthy participants whose tau levels rose the fastest also experienced the steepest increases in depressive symptoms over the follow-up period. In contrast, the rate at which a person’s depressive symptoms worsened was not significantly linked to how quickly they accumulated tau.
A simpler test of whether starting tau levels predicted later changes in mood, or the reverse, did not produce statistically significant results in either group. Next, the team used a mathematical technique called continuous time structural equation modeling, which estimates how one changing variable predicts future values of another variable across uneven time gaps between clinic visits.
Because this method requires a person’s scores to change at least slightly over time, 95 cognitively normal and 26 cognitively impaired participants whose depression scores remained completely flat across all visits were excluded from this specific analysis. Among the remaining 131 cognitively healthy adults, higher tau at a given scan predicted higher depressive symptoms at later visits, though the authors focused on whether this link existed rather than on how its strength changed over different time spans because each person had only a few scans.
Once again, the reverse pathway was not statistically significant: fluctuations in depressive symptoms did not predict later changes in tau. In the cognitively impaired group, neither pathway was statistically significant. To check that their statistical model was working properly with the available sample sizes, the researchers ran the exact same test using objective memory scores instead of mood scores, confirming that higher tau predicted subsequent memory decline in both the healthy and impaired groups.
For their third test, the scientists used an algorithm to estimate the age at which each participant first crossed the threshold into tau positivity in the entorhinal cortex, one of the very first brain areas near the hippocampus to develop tau tangles. Looking at the 62 cognitively normal and 96 cognitively impaired participants who had at least one tau-positive scan, they found that the cognitively impaired group had become tau-positive at an average estimated age of 67.6, about four and a half years earlier than the cognitively normal group’s average onset age of 72.2. Within the cognitively normal group, people who had been tau-positive longer showed steeper increases in depressive symptoms over time, whereas this relationship was not statistically significant in the cognitively impaired group.
“I was particularly surprised that we only observed significant tau-depressive symptom effects in the cognitively unimpaired sample and not in the MCI/AD dementia sample,” Markova said. “I think this really highlights the complexity of what is going on in MCI and AD dementia. To understand the dynamics of these relationships once there is a diagnosis, we may need different, more targeted approaches.”
“This particular study is contributing to a larger body of work demonstrating how late-life changes in mood from your baseline can really be indicative of developing neurodegenerative disease,” Markova explained. “So one major takeaway for a general audience could be about this field of research in general. A significant change in mood or a new neuropsychiatric symptom in later life could offer meaningful information about what’s going on in your brain. Along with complementary information about other biomarkers, tracking changing mood can help us identify who is at an elevated Alzheimer’s disease risk.”
“Also, while the effects in our particular study are small and subtle, we are still able to track significant relationships with tau PET imaging,” Markova added. “This really highlights the power of tau PET imaging.”
The rate-of-change and tau-onset results in the cognitively normal group held after the researchers adjusted for whether participants had amyloid-beta plaques or carried the APOE ε4 gene variant, while the continuous-time models did not include these or other background factors, which the authors left out to keep the models workable. Because about 59 percent of the cognitively normal participants were amyloid-negative at the start of the study, the authors suggest that early tau buildup can contribute to emerging depressive symptoms independently of amyloid plaque pathology.
The researchers propose that tau accumulation in emotion-related brain hubs like the amygdala, or in tiny brainstem clusters that supply mood-regulating chemicals to the rest of the brain, may underlie these subtle shifts in affect. The study itself did not test specific regions, and tau PET scans cannot measure tau in the brainstem.
To see whether concerns about memory loss were influencing the mood scores, the researchers ran an exploratory check by removing one question from the 15-item depression scale: “Do you feel you have more problems with memory than most people?” When they recalculated the models using the remaining 14 questions, the links between tau buildup and worsening depressive symptoms in the cognitively normal group were no longer statistically significant, and the continuous time model could not produce stable estimates. In the cognitively impaired group, slower increases on the 14-item mood score were unexpectedly associated with faster tau buildup.
The authors suggest that removing the memory question may have narrowed the range of scores too much to detect small statistical effects, but they also note that a person’s awareness of their own subtle cognitive decline likely plays a real part in the link between tau and mood. In the authors’ reading, declining mood and subjective memory complaints may represent overlapping expressions of early brain pathology.
There are a few things to keep in mind when interpreting these findings. Because the study is observational, it can show which change tended to come first, not that tau directly causes mood changes. With only about two to three scans per person on average, the timing estimates remain rough, and the effects were small.
Because all participants were screened to exclude clinical depression at the start of the study and mostly reported very mild symptoms, the results cannot rule out the possibility that severe, clinical depression might actively speed up tau spread in a two-way feedback loop, as some animal studies of chronic stress have suggested. In addition, the statistical model used to test directionality over time had to drop more than 40 percent of the cognitively healthy sample because their mood scores never budged, meaning those specific findings apply only to older adults whose mood fluctuates over time.
“One common misinterpretation I want to highlight is that the participants included in these samples are not clinically depressed,” Markova noted. “Our results show that even subclinical changes in mood can significantly track with tau accumulation, and other papers have shown that these subclinical, below-threshold changes can also signify increased risk of going on to develop AD dementia later on. Additionally, the dynamics of these relationships could be very different in samples that have diagnosed depression.”
“As such, one of my long term goals is to follow participants across longer periods of time as some of them transition to a diagnosis of depression and/or a diagnosis of AD dementia,” she said. “This kind of work could give more valuable information about the factors that confer additional risk for or resilience against negative clinical outcomes.”
The lack of statistically significant findings in the mild cognitive impairment and Alzheimer’s dementia group also highlights the difficulty of tracking mood in later stages of disease. The authors note that as memory and thinking decline, self-reported questionnaires become less reliable, and other complex behavioral symptoms or widespread brain changes may overshadow the specific influence of tau. Future studies could combine caregiver ratings with smartphone-based daily mood check-ins or wearable sleep trackers in more racially and ethnically diverse populations, including people with clinical depression, to map out how emotional health and brain biology interact across the full course of aging.
The study, “Tracking the Continuous Time Dynamics of Tau and Depressive Symptoms Across the Alzheimer’s Disease Continuum,” was authored by Teodora Z. Markova, Margo B. Heston, and Anne S. Berry.
