If you are worried about your memory, the most useful thing you can do before an appointment is write things down. It is also the thing most likely to tip into a habit that makes the worry worse.
Both are true, and the difference between them is method.
Why a record is worth having
Memory for memory failures is itself unreliable. You cannot accurately recall how often something happened when the thing in question is recall. Worry makes it worse in both directions: a bad week feels like a decline, and a good week feels like proof it was nothing.
Trajectory is the most informative single piece of evidence, and it is only available if the earlier point was recorded. "About the same, I think" is the answer everyone gives at a six-month review, and it is the one answer nobody can use.
Clinicians act on specifics. Three dated instances of forgetting a conversation is a different consultation from "I feel like it is getting worse".
A record is the only way to see confounders. That the bad days were the four you slept five hours is invisible without notes and obvious with them.
The method: brief, intermittent, instance-based
Log instances, not impressions. Not a rating out of ten. What happened.
14th — could not remember the conversation with my brother on Sunday at all. He had to remind me of the whole thing. 17th — lost keys, found in coat. Normal for me. 19th — asked the same question twice in one afternoon, noticed by my wife.
Three lines. Two minutes a week.
Include the ordinary ones. Recording only the frightening instances produces a document that looks worse than your life, which is bad for your anxiety and misleading to a clinician. Noting "lost keys, normal for me" is doing real work.
Note the obvious confounders on bad days. Sleep, alcohol, whether you were unwell, whether it was an unusually demanding week, any medication change. Four fields at most.
Write on the day, or not at all. A note reconstructed on Sunday about Tuesday is a memory of a memory failure.
Do not log every day. This is the most important rule. Log when something happens. A daily practice of assessing your own cognition is a practice of scanning for decline, and it will reliably find some.
The trap
There is a version of this that becomes continuous self-monitoring: assessing every hesitation, logging every misplaced object, comparing today against yesterday. It produces a thick document and a much worse month.
Two reasons it backfires. Attention makes things more salient — watch closely for word-finding pauses and you will find them, because they were always there. And anxiety independently impairs memory and word retrieval, so vigilance measurably worsens the thing being measured.
If you notice any of the following, stop and switch to a fixed weekly review:
- Checking whether your memory is working today
- Testing yourself
- Feeling worse after writing
- Rereading old entries to compare
- Logging things you would not have noticed a month ago
A once-weekly ten-minute review of what actually happened captures everything a clinician needs and does not involve living inside the question.
Ask someone else to keep one too
Higher value than your own, and underused. Ask a partner or someone who sees you daily: would you note down anything you notice, with the date?
Their record is corroboration, which is weighted heavily in any assessment, and it catches what you cannot see. Two months of theirs alongside two months of yours is a genuinely strong basis for a conversation.
What to bring to the appointment
Not the raw log. A summary you wrote from it:
Over eight weeks: three instances of forgetting a conversation entirely. Daily misplacing of keys and phone, but that has always been true. Two occasions of repeating a question, both noticed by my wife. The worst week was the one I slept badly, around the 12th. No change in medication. My wife's note is attached.
That paragraph took four minutes to write and does more than an hour of talking. Bring the log as backup in case of detailed questions.
For a six-month review
If the plan is to reassess later, this is when a record pays for itself. Keep the same format, in the same place, and put a reminder in the calendar for a week before the appointment to write the summary.
The comparison — "at the start it was roughly once a fortnight, it is now roughly twice a week" — is exactly the evidence a trajectory-based assessment needs and almost never gets.
The honest limit
A record shows what happened and what co-occurred. It cannot tell you why, and it cannot tell you what it means. Two things moving together in your notes is an association — a good reason to look, never a conclusion on its own.
That restraint is what keeps a record useful. Its job is to make the first question a better question, and then to hand the answering to someone qualified to do it.