Table 2.

HDTwin text processing rules. These results are used to process information found in text including journal entries and interview assessments.

Marker set	Rule
Speech	If journal_text is empty the more likely mild cognitive impairment If journal_text has a large vocabulary, long sentences, and/or high sentence complexity then more likely healthy If journal_text has a small vocabulary, short sentences, and/or low sentence complexity then more likely mild cognitive impairment If journal_text uses positive emotion words then more likely healthy If journal_text uses negative emotion words then more likely mild cognitive impairment If journal_text appears to have more than 1 entry then more likely healthy If journal_text appears to only have 1 entry then more likely mild cognitive impairment
Baseline testing session	if some interview_assessment ratings are <= 3 then more likely mild cognitive impairment if most interview_assessment ratings are >= 4 then more likely healthy if the interview_assessment explanations suggest the participant confidently answered the questions correctly then more likely healthy if the interview_assessment explanations suggest the participant struggled to answer the questions correctly then more likely mild cognitive impairment