Personal Sensing: Understanding Mental Health Using Ubiquitous Sensors and Machine Learning

20 important questions on Personal Sensing: Understanding Mental Health Using Ubiquitous Sensors and Machine Learning

What are examples of high-level behavioral markers? (5)

1. Hedonic activity
2. Fatigue
3. Depressed mood
4. Social avoidance
5. Stress

What is personal sensing?

The utilization of sensor data to estimate behaviours

What is the goal of personal sensing?

To convert the potentially large amount of raw sensor data into meaningful information regarding behaviors, thoughts, emotions, and clinical states and disorders.
  • Higher grades + faster learning
  • Never study anything twice
  • 100% sure, 100% understanding
Discover Study Smart

What is a behavioral marker?

Behaviors, thoughts, feelings, traits or states identifies using personal sensing

What are the steps taken in the hierarchical model leading from data to knowledge? (4)

1. Framework captures raw sensor data
2. Data is converted into feature
3. Behavioral markers are defined (through machine learning)
4. A set of features and markers is used to identify clinical states

How are high-level behavioral markers commonly developed?

Using machine learning and data-mining methods to uncover which features and sensor data are useful in detecting the marker.

What is a shortcoming of personal sensing?

That some symbols might simply not be detectable. Besides, personal sensing may uncover other predictors that have not been considered up to date.

How is non-workday sleep used in judging someone's sleep?

Non-workday sleep can be used to estimate a person's chronotype. Changes in sleep patterns and non-workdays can in that way identify social jet lag.

What are the current challenges in personal sensing? (6)

1. Study quality and reproducibility (difference in study design & questionable sources receiving similar attention)
2. The curse of variability (differences in phone usage)
3. The unknown expiration date (algorithms increasingly inaccurate & short shelf-life)
4. Balancing accuracy and visibility
5. The certainty of uncertainty (what is acceptable)
6. Privacy, ethics, and the naked truth (de-identification, as much control as possible)

What is the meaning of privacy, security, trust, and value? (4)

1. Privacy: people have choice and control over the use of their own data
2. Security: protections to ensure people's choices are followed
3. Trust: data will be used appropriately
4. Value: benefit due to use of data

What are the potential applications of personal sensing? (2)

1. Integration into existing models of care
2. Behavioral intervention technologies

In what ways can mental healthcare be enhanced by the use of personal sensing? (3)

1. Identifying people in need of treatment
2. Accelerating access to treatment
3. Monitoring functioning during or after treatment

In what way could behavioral intervention technologies be improved?

Success will be more likely if what is sensed, and how sensed data are used, speak to the user's personal goals.

What is the advantage of the fact that we use sensors in our daily lives?

Personal sensing offers the potential to measure human behavior continuously, objectively, and with minimal effort from the user.

How can raw sensor data be translated into knowledge?

By using a layered, hierarchical approach in which sensor data are converted into features, and features are combined to estimate behaviors, moods, and clinical states

What is a finding of a growing number of studies on the use of phone sensor data?

That it can, by use of machine learning, provide markers of sleep, social context, mood, and stress.

What was found about depression and mood states in bipolar disorder?

That they can be estimated using a variety of phone sensor data. GPS features measuring entropy and the circadian rhythm have been correlated with depression. Posts on social media can also identify people who are depressed or likely to become depressed.

What is the main limitation regarding the evidence on the use of phone sensor data?

That most studies were small, on the order of 7-30 participants, who are mostly college students. There is little evidence to support replicability.

In what ways do machine-learning methods vary?

Some rely on user-generated labels, while others uncover patterns in unlabeled data. It however appears that labeling often improves and helps algorithms adapt to new circumstances.

What is the main conclusion about personal sensing?

That it is feasible to obtain data from personal sensing using everyday sensors. However, numerous challenges must be overcome before it is viable for clinical deployment.

The question on the page originate from the summary of the following study material:

  • A unique study and practice tool
  • Never study anything twice again
  • Get the grades you hope for
  • 100% sure, 100% understanding
Remember faster, study better. Scientifically proven.
Trustpilot Logo