Depression is one of the few psychiatric syndromes with genuinely observable signs, and the sign-based tradition — Widlöcher's work on retardation, Parker's on melancholia as a disorder of movement — argues they are the most specific thing about melancholic depression. This scheme brings that tradition to a video timeline: nine signs, each rated 0 to 3, judged only from what the recording shows.
Rate what you see, not what you are told. That single rule is what distinguishes this from every symptom scale in the text library. A patient who describes crushing slowness but moves and speaks briskly on camera scores low here — and that discrepancy is itself a finding, not an error to smooth over. HAM-D and MADRS already capture the report; this captures the sign.
Two families, rated differently. Retardation signs are stable states, so rate their average across the interview and, where two ratings look equally defensible, take the less severe. Agitation signs come in episodes, so mark the episodes you saw and let severity reflect their typical intensity rather than the single worst second. Both conventions exist because raters left to their own devices drift in opposite directions on the two families.
Give it enough tape. These are judgements about a person's manner over time, not snapshots. The sign-based literature works from interviews of twenty minutes or more, and ratings from a two-minute clip should be labelled as such.
Stereotypy depends on agitation. Repetitive movements — hand-wringing, picking, rocking — only count as agitated stereotypy in the presence of motor agitation. If you rated no agitation, do not rate stereotypy either; otherwise ordinary habitual movement gets scored as pathology.
What is missing on purpose. Psychomotor change also shows up in speech timing — response latency, pausing, slowed articulation — which needs the audio, not the picture. That lives in the Psychomotor Speech Signs scale in the audio library, and because audio scales can be applied to video studies you can run both against the same recording and get the full picture.
If your protocol specifically requires the CORE measure, license it from Cambridge University Press; this scheme covers the same construct in Tagaroo's own words and is not a substitute for that instrument in a study that names it.