Innovations in Clinical Neuropsychology Summary
Week 1.
Lecture
- Many apps for brain training claim: improve performance in school/work. Delay age related
cognitive decline. Reduce impairment from health conditions.
It can improve cognition, but often not generalizable
Little evidence though
- Near transfer effects vs far transfer effects
- Technology and development: passive vs active vs interactive
- Technology and neuropsychology:
6% of tools are computerized obsolete technologies (no new ideas)
- Why should we innovate:
Technical development: it is available so lets use it
Scientific motivation: more insights into cognitive function
Less limitations than existing materials: more data, and maybe less susceptible to
issues
- Innovation categories: diagnostic tools, treatment tools, ehealth, neurotechnology, non
technological innovations
- Innovations in diagnostics:
Computer based assessment of cognition mostly used in military and sports, less
used in clinical practise due to financial costs, lack of normative data, concerns about
utility and validity
Digital versions of existing standardized material
- Randolph criteria:
Test-retest reliability
Sensitivity: those who are correctly given a diagnosis: true positives
Validity: does it measure what it is supposed to
Reliable change scores: to test change over time: difference score of 2 tests of same
person over time
Clinical utility
- Recommendations for healthcare professionals for NCATs: use cautiously (not standalone),
use what feels best to fit their needs and targeted population, consider age and education as
context, remain aware of forthcoming recommendations
- Advantages of computerized NCATs:
More detailed measurements: time (eg. initiation, inspection, per item: easier to
see/set time), drawing and writing (eg. start, clustering, neglect: computer can keep
score automatically and better)
Tailoring to specific needs is possible
Ease of use
Reducing human error
Mimicking everyday situations (measuring everyday functioning)
Remote and portable testing
- Disadvantages of NCATs
Norm data not directly transferable: validity and reliability needs to be proven still
Technical requirements
, Training of clinicians
Cognitive processes possibly different for digital environment
Cybersickness
Novelty costs (adjustment time needed)
Privacy issues (data storage)
- Ecological validity of VR diagnostics:
Conditions affects performance selectively
Conditions can be EG:
1. Locomotion: real life
2. Hybrid: locomotion + table
3. Virtual: no locomotion
4. Virtual +: no locomotion + you have a compass
Landmark and route knowledge were unaffected in the conditions
Locomotion benefits survey knowledge
- Treatment tools with innovation:
Less developed than diagnostics
Rapidly increasing
Studies focus mostly on physical therapy
Increased ecological validity (but see diagnostics)
- Compensation= use of other function
Other function reduces impairment (eg. changing cognitive strategy)
Situations can be presented in which alternative strategy is stimulated
- Restoration= improve lost/ reduced function
Eg. memory training
Relevant situations can be presented that stimulate function use
- Ethical considerations: real vs virtual. Social interactions?. Level of understanding of patients
- Ehealth:
Emerging field
Health services and information delivered/ enhanced through internet and related
technologies
Tracking and monitoring: additional info, smartphone/ wearables, data analysis
Records: other form of archiving, easier communication (between clinician and
patient), privacy issues
Communication: personal
contact in other form,
providing additional info,
smartphones and online
portals
Self management: addition to
regular care, eg. wearables,
direct link to clinician
Goals: improve access to
health care recourses, educate
and support patients and
caregivers, empower patients
to become active in disease
management and treatment,
improve understanding of disease (progression)
, - Brain compute interface: BCI (See picture)
- Non technical innovations:
ICF model: international classification of functioning, disability and health. = classify
health. Complement to ICD 10 (classify causes of death/ disorder)
Biopsychosocial model: principles: universality, parity, neutrality, environmental
factors
Includes context
RDoC framework: research domain criteria: based on function not symptoms
(transdiagnostic). Accounting for environment (eg. culture, social determinants of
health). Multi level approach (measurement and integration of diverse data)
RDOC matrix: tool to help implement the principles of RDoC
Physiology: measures that are well established constructs, but don’t necessarily tap
circuits directly (eg. hearth rate, cortisol)
Behaviour: behavioural tasks (eg. working memory) or systematic behavioural
observations
- Performance monitoring= ability to monitor and adjust ongoing performance
- Combination of methods (multi level) increases clinical utility
- Neurophysiological correlates (EEG):
ERN: error related negativity
N2 and N450: response inhibition
Article: Technology crisis
- Often now use data collection methods that are slow, inefficient, expensive
- First more brain behaviour relations were studied.
, - Modern CNP emerged in 1960s-1970s (specialty journals and professional organizations)
- Transition from neuropsychology 1 to 2.0 marked by shift from isolated brain function
measures to psychometrically validated tests
- CNP often uses tests first used in other purposes
- Measures often applied to certain settings: student classification, military assessment,
immigration process
- Boston process approach: emerged due to limitations in existing tests. Focus on qualitative
aspects of test performance for neurological and psychiatric conditions
- Neuropsychology lags behind a bit, relies still a lot on paper and pencil tasks. The ties to
traditional test publishing companies hinders technological innovation
- Often traditional tests are preferred over innovation by test publishers
- Many tests minimally revised (eg. WAIS-IV, verbal learning assessment (use old word lists))
- Limited innovation in compute tests
- Insufficient psychometric studies and poor quality of normative data for computerized tests
- Only 6% of tests were computerized when computers became a thing, but they were used
little
- Reasons for low computer adoption: high costs of tests, lack of normative data, concerns
about utility and validity
- Boston process approach is made to align traditional tests with clinical needs
- Combining technology driven assessment and passive data collection can offer complete
profiles of cognition and behaviour
- Digitalization can improve: administration, automation in scoring and administration
- Technology can help measure speed, accuracy, pauses, consistency in responses
- Smart pencils can be used for tasks to record and analyse pressure/ tremor/ response
patterns (eg. trail making test, clock drawing task)
- Visuospatial construction tasks can use sensor based tangible stimuli to record and automate
scoring and response pattern analysis
- Verbal fluency tasks can use speech recognition for automatic scoring and response pattern
analysis
- Verbal memory test can use speech recognition with recordings
- Nonverbal tasks can use multimodal paradigms
- Automated methods can help accuracy and consistency in data collection and minimize
administrator input
- Digital assessments ensures: standardized instructions/ demonstrations, reduces
administrator errors, improves reliability, minimise errors in scoring/ data entry, saves time,
enhances clinical efficiency, faster results (also improves access to CNP services)
- Computer adaptive testing: CAT: adapts difficulty based on patient performance, improving
precision and requiring fewer items to determine ability
- Item response theory: IRT: measures laten variables. Enables longitudinal monitoring, test
security
- Objective test item calibration supports cultural diverse application and reduced need for
large scale test revision
- Challenges of using computers: rigid systems hinders individual flexibility, behavioural
impulsivity might lead to too fast answers without ability to revise them, privacy concerns
- You need informed consent for cloud based data storage
- Portable devices enable remote testing, maintain consistency, integrate other technologies,
track progress
Week 1.
Lecture
- Many apps for brain training claim: improve performance in school/work. Delay age related
cognitive decline. Reduce impairment from health conditions.
It can improve cognition, but often not generalizable
Little evidence though
- Near transfer effects vs far transfer effects
- Technology and development: passive vs active vs interactive
- Technology and neuropsychology:
6% of tools are computerized obsolete technologies (no new ideas)
- Why should we innovate:
Technical development: it is available so lets use it
Scientific motivation: more insights into cognitive function
Less limitations than existing materials: more data, and maybe less susceptible to
issues
- Innovation categories: diagnostic tools, treatment tools, ehealth, neurotechnology, non
technological innovations
- Innovations in diagnostics:
Computer based assessment of cognition mostly used in military and sports, less
used in clinical practise due to financial costs, lack of normative data, concerns about
utility and validity
Digital versions of existing standardized material
- Randolph criteria:
Test-retest reliability
Sensitivity: those who are correctly given a diagnosis: true positives
Validity: does it measure what it is supposed to
Reliable change scores: to test change over time: difference score of 2 tests of same
person over time
Clinical utility
- Recommendations for healthcare professionals for NCATs: use cautiously (not standalone),
use what feels best to fit their needs and targeted population, consider age and education as
context, remain aware of forthcoming recommendations
- Advantages of computerized NCATs:
More detailed measurements: time (eg. initiation, inspection, per item: easier to
see/set time), drawing and writing (eg. start, clustering, neglect: computer can keep
score automatically and better)
Tailoring to specific needs is possible
Ease of use
Reducing human error
Mimicking everyday situations (measuring everyday functioning)
Remote and portable testing
- Disadvantages of NCATs
Norm data not directly transferable: validity and reliability needs to be proven still
Technical requirements
, Training of clinicians
Cognitive processes possibly different for digital environment
Cybersickness
Novelty costs (adjustment time needed)
Privacy issues (data storage)
- Ecological validity of VR diagnostics:
Conditions affects performance selectively
Conditions can be EG:
1. Locomotion: real life
2. Hybrid: locomotion + table
3. Virtual: no locomotion
4. Virtual +: no locomotion + you have a compass
Landmark and route knowledge were unaffected in the conditions
Locomotion benefits survey knowledge
- Treatment tools with innovation:
Less developed than diagnostics
Rapidly increasing
Studies focus mostly on physical therapy
Increased ecological validity (but see diagnostics)
- Compensation= use of other function
Other function reduces impairment (eg. changing cognitive strategy)
Situations can be presented in which alternative strategy is stimulated
- Restoration= improve lost/ reduced function
Eg. memory training
Relevant situations can be presented that stimulate function use
- Ethical considerations: real vs virtual. Social interactions?. Level of understanding of patients
- Ehealth:
Emerging field
Health services and information delivered/ enhanced through internet and related
technologies
Tracking and monitoring: additional info, smartphone/ wearables, data analysis
Records: other form of archiving, easier communication (between clinician and
patient), privacy issues
Communication: personal
contact in other form,
providing additional info,
smartphones and online
portals
Self management: addition to
regular care, eg. wearables,
direct link to clinician
Goals: improve access to
health care recourses, educate
and support patients and
caregivers, empower patients
to become active in disease
management and treatment,
improve understanding of disease (progression)
, - Brain compute interface: BCI (See picture)
- Non technical innovations:
ICF model: international classification of functioning, disability and health. = classify
health. Complement to ICD 10 (classify causes of death/ disorder)
Biopsychosocial model: principles: universality, parity, neutrality, environmental
factors
Includes context
RDoC framework: research domain criteria: based on function not symptoms
(transdiagnostic). Accounting for environment (eg. culture, social determinants of
health). Multi level approach (measurement and integration of diverse data)
RDOC matrix: tool to help implement the principles of RDoC
Physiology: measures that are well established constructs, but don’t necessarily tap
circuits directly (eg. hearth rate, cortisol)
Behaviour: behavioural tasks (eg. working memory) or systematic behavioural
observations
- Performance monitoring= ability to monitor and adjust ongoing performance
- Combination of methods (multi level) increases clinical utility
- Neurophysiological correlates (EEG):
ERN: error related negativity
N2 and N450: response inhibition
Article: Technology crisis
- Often now use data collection methods that are slow, inefficient, expensive
- First more brain behaviour relations were studied.
, - Modern CNP emerged in 1960s-1970s (specialty journals and professional organizations)
- Transition from neuropsychology 1 to 2.0 marked by shift from isolated brain function
measures to psychometrically validated tests
- CNP often uses tests first used in other purposes
- Measures often applied to certain settings: student classification, military assessment,
immigration process
- Boston process approach: emerged due to limitations in existing tests. Focus on qualitative
aspects of test performance for neurological and psychiatric conditions
- Neuropsychology lags behind a bit, relies still a lot on paper and pencil tasks. The ties to
traditional test publishing companies hinders technological innovation
- Often traditional tests are preferred over innovation by test publishers
- Many tests minimally revised (eg. WAIS-IV, verbal learning assessment (use old word lists))
- Limited innovation in compute tests
- Insufficient psychometric studies and poor quality of normative data for computerized tests
- Only 6% of tests were computerized when computers became a thing, but they were used
little
- Reasons for low computer adoption: high costs of tests, lack of normative data, concerns
about utility and validity
- Boston process approach is made to align traditional tests with clinical needs
- Combining technology driven assessment and passive data collection can offer complete
profiles of cognition and behaviour
- Digitalization can improve: administration, automation in scoring and administration
- Technology can help measure speed, accuracy, pauses, consistency in responses
- Smart pencils can be used for tasks to record and analyse pressure/ tremor/ response
patterns (eg. trail making test, clock drawing task)
- Visuospatial construction tasks can use sensor based tangible stimuli to record and automate
scoring and response pattern analysis
- Verbal fluency tasks can use speech recognition for automatic scoring and response pattern
analysis
- Verbal memory test can use speech recognition with recordings
- Nonverbal tasks can use multimodal paradigms
- Automated methods can help accuracy and consistency in data collection and minimize
administrator input
- Digital assessments ensures: standardized instructions/ demonstrations, reduces
administrator errors, improves reliability, minimise errors in scoring/ data entry, saves time,
enhances clinical efficiency, faster results (also improves access to CNP services)
- Computer adaptive testing: CAT: adapts difficulty based on patient performance, improving
precision and requiring fewer items to determine ability
- Item response theory: IRT: measures laten variables. Enables longitudinal monitoring, test
security
- Objective test item calibration supports cultural diverse application and reduced need for
large scale test revision
- Challenges of using computers: rigid systems hinders individual flexibility, behavioural
impulsivity might lead to too fast answers without ability to revise them, privacy concerns
- You need informed consent for cloud based data storage
- Portable devices enable remote testing, maintain consistency, integrate other technologies,
track progress
