Guide-linked quiz · Medium Level
Digital Phenotyping and Mood Tracking Quiz
Interpret smartphone mood signals cautiously and account for indirect measures, missing data, privacy, consent and limits on diagnosis.
Question 1 of 8
How does digital phenotyping use personal-device data in mental health research?
Preview every question8 questions
Answers and explanations appear as you take the quiz.
- How does digital phenotyping use personal-device data in mental health research?
- A one-time questionnaire with no device-derived behavioral data
- A single emergency screening interview conducted by phone
- Using device-derived behavior or physiology to study patterns over time
- An electronic health-record billing analysis
- Why is low smartphone activity not a direct measure of depression?
- Phones record mood directly
- Many contextual factors can change phone use
- Low activity always indicates illness
- Sensor data are immune to missingness
- What can an observational group comparison establish?
- A causal mechanism for every participant
- That a model is clinically valid everywhere
- That sensor behavior has no confounders
- Patterns associated with groups in that study, not an individual diagnosis or cause
- Why can missing sensor data bias a mood model?
- Missingness may relate to device use or life context
- Missing values always indicate healthy mood
- All phones record every sensor identically
- A model can reconstruct all missing context exactly
- What privacy risk can continuous sensor collection create?
- It cannot identify anyone if names are removed
- Derived features contain no private information
- It can reveal routines, relationships and sensitive events
- Sensor data are always public
- Which evaluation split helps test whether a model transfers to new participants?
- Randomly duplicate overlapping windows across splits
- Hold out participants or time periods where appropriate
- Evaluate only on the training data
- Choose the split after seeing final results
- How should a personal mood-tracking app present model output?
- As a definitive clinical diagnosis
- As an automatic treatment plan
- As proof that the user has a disorder
- As uncertain context for reflection, not a diagnosis
- What should participants be able to understand about passive sensing?
- What is collected, how it is used, retained, shared and deleted
- Only the app's name
- Only the model architecture
- That collection cannot be paused
Keep testing yourself
More quizzes picked for your level.