In a 2015 study of 335 adults, Kamen Tsvetanov and his colleagues found that the parts of the brain involved in seeing and hearing showed smaller fMRI responses to the task at older ages. Then the team reanalysed the scans, using each participant’s resting scan to account for differences in blood vessel function. Much of the apparent age difference fell away. The volunteers had seen the same patterns and heard the same tones. What changed was how the researchers interpreted the signal.
A chequerboard and a tone
A button press tells a researcher more when they know what the participant saw or heard just before it. In a round-based game at an online casino such as scizino, a player acts and then sees the outcome. A researcher studying anticipation would have to record those two moments separately. The Cam-CAN experiment had a simpler aim. Its timing centred on when a sight or sound appeared and when the person in the scanner detected it.
On most trials, volunteers saw a black-and-white chequerboard and heard a tone. They pressed a button with their right index finger when they detected either cue. Some trials offered only a sound or only a visual pattern. Between events there were periods with just a fixation cross. The task gave the team a way to look at responses in areas involved in vision, hearing and movement within the same experiment.
What each instrument could tell them
Functional MRI, or fMRI, tracks a signal that depends on how much oxygen the blood carries. It’s usually called BOLD, short for blood oxygenation level-dependent. When neurons become active, local blood flow tends to change. That link is what makes BOLD useful. It also means that a blood vessel that responds differently can change the signal even if the underlying activity is similar. The study brought together several measurements to get a handle on that problem.
| Record | What it captured | Role in the study |
|---|---|---|
| Task fMRI | BOLD changes during the cues | Estimate the response to the task |
| Resting fMRI | BOLD fluctuations without the task | Estimate a scale for each scan location |
| Resting MEG | Magnetic signals from brain activity | Compare variation in neural activity |
| Heart recordings | Rate and variation between beats | Provide clues about cardiovascular function |
MEG, short for magnetoencephalography, records faint magnetic fields generated by electrical activity in the brain. Those recordings gave the team a way to examine neural activity at rest without relying on the blood oxygen signal. The heart recordings helped them ask a separate question about the circulation. With both records, they could examine which factors were associated with the ups and downs of BOLD while volunteers rested in the scanner.
The comparison that changed the picture
Look at Figure 4 in the paper and you’ll see two versions of the age comparison. In one, the team used the task responses as estimated from fMRI. In the other, it scaled those estimates by the amount of fluctuation at rest, using the matching location in each person’s brain scan. The idea was to allow for the fact that the size of a BOLD response depends partly on how that person’s circulation behaves.
The age differences in visual regions and much of the auditory cortex became weaker after scaling. A finding in the movement system survived the same adjustment. The paper’s discussion puts the result this way.
“The effect of age in the motor cortex remained significant.”
That detail matters because a headline about differences disappearing would miss part of the result. It also gives readers a more useful way to read the figure. Follow the same brain region across the two analyses and see which patterns survive the adjustment.
| Where to look | Before scaling | After scaling |
|---|---|---|
| Visual areas | Smaller responses at older ages | The age difference weakened |
| Auditory areas | Smaller responses at older ages | Much of the difference weakened |
| Motor area on the side of the responding hand | Larger responses at older ages | The association with age persisted |
| Visual and auditory areas together | Estimates from task scans alone | Direct comparison showed a smaller age effect |
What happens between a scan and a figure
Life as a neuroscience researcher often involves working through the records that go into a single figure. In this study, there were scans taken at rest, scans taken during a task and recordings from another instrument. A good set of notes helps a colleague trace any plotted point back to the session and understand how it got there. In a neuroscience laboratory, that might mean keeping track of a few ordinary details.
- Which session a scan or recording came from
- When the cues appeared and which responses were registered
- Where movement or a recording problem affected the data
- How the resting and task measurements were matched
- Which records were excluded from a comparison and why
How to read the next ageing study
Dr Kamen Tsvetanov and his co-authors return to the problem in a later review of fMRI and ageing. The review asks what we can learn by studying changes in the vessels themselves. The authors describe vascular ageing in these terms.
“an important feature in its own right, to be assessed alongside measures of neuronal ageing.”
When a scan appears in a research story, the methods and figure captions are good places to start. These questions apply to the papers listed under publications and the talks listed under presentations.
- What did volunteers do while the signal was recorded?
- What did the instrument measure directly?
- Did the study collect separate information on blood vessels?
- Which results persisted when the researchers adjusted for those differences?
- Did the team compare people of different ages or follow the same people over time?
That last question is easy to miss. A comparison across ages tells you how people in the study differed at the time they were tested. Tracking change within a person takes repeat measurements. The 2015 paper is useful for another reason. It gives you two analyses of the same task to compare, so you can see what happens to an apparent age difference when the researchers change how they read the signal.