Changes in neuronal activation patterns in response to androgen deprivation therapya pilot study

He's done this parking garage a hundred times. Same ramp, same turn, same mental snapshot of where he left the car. But today, a few months into his prostate cancer treatment, something is slightly off. The map won't quite hold in his mind. He circles. He doubts himself. He finds the car eventually. But something has shifted. That experience — vague, easy to dismiss, hard to measure — is exactly what Cherrier and colleagues decided to examine directly. They put five men inside an MRI scanner before and after androgen deprivation therapy and watched, in real time, as specific brain regions went quiet. Prostate cancer is the most commonly diagnosed non-skin cancer in American men, with over 218,000 new cases identified each year. Androgen deprivation therapy, known as ADT, works by reducing circulating androgens like testosterone. It was once reserved for metastatic disease but is now prescribed much more broadly: for localized disease in combination with radiation, and for roughly 22 percent of patients who fail primary therapy and develop a rising prostate-specific antigen without detectable spread. Many of these men have no disease symptoms and can live for years on ADT. That's the double-edged reality. The treatment works, but it comes with costs. Those costs include the well-documented physical ones: bone density loss, muscle atrophy, anemia, and increased cardiovascular risk. Cognitive complaints are also real. Men undergoing ADT report difficulties with verbal memory, spatial reasoning, and attention. The problem is that patient reports and clinical intuition are not the same as biological evidence. Cherrier and colleagues wanted to know what was actually happening inside the brain, not just what patients described afterward. Their method was functional magnetic resonance imaging, or fMRI. The technique measures blood oxygen level-dependent signal as a proxy for neural activity. When a brain region is active, blood flow there increases, and the scanner detects that change. It's not a perfect window, but it's the clearest one we have into working brain function in living people. The design was straightforward and purposeful. Twelve prostate cancer patients were recruited, and five completed usable scans across both timepoints. Seven age and education-matched healthy controls completed the same protocol. Everyone was scanned twice — once before ADT began and again after nine months of treatment. At each session, participants performed three visuospatial tasks: an environmental memory encoding run, a spatial recognition run, and a mental rotation task using three-dimensional Shephard-Metzler figures. The researchers used independent components analysis to strip out motion and artifact from the imaging data, then applied general linear models with random field theory correction to identify regions showing changed activation over time. Before looking at brains, the team confirmed the treatment was working hormonally. ADT participants' total serum testosterone dropped from an average of 3.52 nanograms per milliliter at baseline to just 0.25 nanograms per milliliter at month nine. Controls stayed flat, moving from 3.8 to 4.3 nanograms per milliliter. The group effect and the group-by-time interaction were both highly significant, with F statistics above 37 and 53 respectively. The androgens were gone. The question was what that absence had done upstream, in the brain. Here is what the scans showed. In the ADT group, task-related blood oxygen level-dependent activation was reduced at month nine compared to baseline in a single, localized cluster in the right parietal-occipital region — specifically the precuneus, cuneus, and lateral occipital lobe. The overall cluster had a corrected p-value of 0.011, covered 4.72 cubic centimeters, and peaked at a Z-score of 4.15. No comparable reductions appeared in the control group. And critically, there were no differences between groups at baseline — both groups looked essentially the same before treatment began. That matters for causal inference. The divergence appeared after ADT. Breaking it down by task clarifies where the effect was sharpest. During the spatial recognition run — where participants judged whether objects were correctly positioned or rearranged — the ADT group showed a strong baseline-over-month-nine effect in the right cuneus and precuneus, with a cluster p-value of 0.001 and a Z-max of 4.88. The between-group difference in change over time reached a p-value of 0.032. During the mental rotation task, a similar pattern emerged: ADT participants showed decreased activation in the right cuneus and precuneus at month nine, with a cluster p-value of 0.002 and a Z-max of 5.67, with the between-group change again significant at a p-value of 0.031. Both the spatial recognition and the mental rotation tasks showed ADT-specific reductions. The encoding task showed decreased activation in both groups — not distinctive to ADT. Now here is the part that really demands attention. Behavioral performance on these tasks did not decline. ADT participants scored thirty-six out of sixty-four possible points on the environmental memory task at baseline and thirty-five point seventy-five at month nine — essentially identical. Controls moved from thirty-one point three to thirty-eight. No significant behavioral decline for either group. The brain was doing something different, but the output looked the same. What does that mean? Cherrier and colleagues suggest two possibilities worth considering. One is that the neural signal decreased before behavior degraded — that the imaging caught an early change that a behavioral test couldn't yet detect. The other is that the task wasn't sensitive enough to reveal performance differences in a sample this small. Either way, the brain was not doing the same thing it had been doing nine months earlier. The blood oxygen level-dependent signal changes in individuals ranged from roughly zero point twenty-five to one point zero percent of total signal — small, but the study's design had greater than eighty percent power to detect changes around zero point seventy-five percent. These were real changes, not noise. Why the parietal-occipital region specifically? This is the cortex you recruit when you construct a mental map, rotate an object in your mind, or remember where something was in space relative to everything else. It's the machinery behind reading a map, reassembling furniture from a diagram, and navigating an unfamiliar parking garage. Cherrier and colleagues connect these laboratory tasks directly to those everyday functions — and that connection is not incidental. If ADT is selectively reducing activation in the system that handles three-dimensional spatial transformation, the implications extend well beyond a scanner. The paper situates this finding within a broader literature linking androgens to regional brain activity. Prior imaging work using positron emission tomography had shown androgen-linked changes in regional metabolism and blood flow. This study adds direct fMRI evidence that task-related activation in the parietal-occipital cortex specifically declines during ADT, and that the decline is regionally selective, not a global flattening of brain response. The limitations are real, and the authors state them plainly. Five ADT patients is a small sample. The participants were highly educated — averaging seventeen years of schooling — which may have buffered behavioral performance through what researchers call cognitive reserve. The study offers no comprehensive neuropsychological battery alongside the imaging. It is, as Cherrier and colleagues call it, a pilot. But the strengths matter too. Prospective enrollment before treatment began. Within-subject repeated measures, so each participant served as their own control. No baseline differences between groups. Confirmed hormonal suppression. A pre-registered comparison of specific regions tied to the tasks performed. These design choices give the finding traction it wouldn't have from a cross-sectional study of men already on ADT. ADT extends lives. That is not in question. But these findings raise something worth carrying forward: what if an effective cancer treatment is also quietly reducing the neural activation that supports spatial cognition — before patients notice, before behavioral tests catch it? Scaled-up neuroimaging work, replicated in larger and more diverse cohorts, could eventually inform decisions about treatment duration, monitoring protocols, or neuroprotective interventions. None of that follows automatically from one pilot study. What does follow is a clear, testable hypothesis, grounded in a specific brain region, a specific cognitive domain, and a measurable biological mechanism. That is what good preliminary work looks like. The honest answer, for now, is that we've watched the signal dim. We don't yet know exactly what that costs. This lecture was created by ennepō. Go to https://ennepo.ai to Discover, Create and Follow the latest research in your field. Read when you can. Listen when you want to.