Most of the variation families notice between one session and the next has nothing to do with the software. It has to do with the ordinary conditions the session happened in. This post lists the ones that come up most often, so that a week that looks like a step backwards can be read for what it usually is.
Sleep
A short night shows up in an EEG recording more reliably than almost anything else. A session run on four hours of sleep is not comparable with one run on nine, and it is not worth drawing conclusions from. If the week has been short on sleep, the honest move is to note that next to the record rather than to read the numbers as progress or regress.
Time of day
Alertness follows a daily curve, and the curve differs from person to person. Some children settle best straight after school, others not until after a meal and a break. Picking one slot and keeping it makes the record far easier to interpret, because the time of day stops being a variable.
Food and water
A session attempted while someone is hungry or thirsty tends to be a short and unhappy one. Nothing elaborate is needed; a normal meal or snack beforehand and a glass of water is enough. We make no claims about diet and brain activity, only the practical point that an uncomfortable child does not sit still.
Sensor contact
A headband that has slipped, hair trapped under a sensor, or dry skin will all degrade the signal. If a recording looks unlike every other recording, the first thing to check is the fit, not the child.
The room
A quiet corner with the door shut and phones elsewhere is worth more than any setting in the app. Interruptions do not just break concentration; they show up in the recording.
Illness and hard days
If a child is unwell, upset, or has had a difficult day at school, skipping the session is the better decision. There is nothing to be gained from a recording made under those conditions, and forcing one tends to make the next few harder.
What we do not claim
Auto Train Brain is not a medical device. It is not intended to diagnose, treat, cure or prevent any disease or condition, and nothing above should be read as health advice. Published research on the approach is listed on our research articles page; those studies were small and their results should be read as preliminary.