Skip to content
Contact
☰
Color scheme: Dark.
Back to research
RESEARCH SPACE G

Human–Machine Interfaces & Neuroadaptive Systems

How do interfaces and adaptive systems influence perception, behavior, cognitive work and human–AI interaction, and how could systems respond in future to measured or inferred human states?

What this space studies

How interfaces and adaptive systems influence what people perceive and decide and how they work with AI, and how a system could adapt to changing states in future. An interface also changes the conditions of later interaction, and this feedback is part of the study. Today interaction data and behavioral proxies exist for this. Physiological signals, wearables, EEG, BCI and a full adaptive response are planned and not implemented.

Central questions

  • How do interfaces shape decisions, attention and system understanding?
  • Which states can be estimated from behavior, and where does that end?
  • What must a system show that adapts to human states?

Interfaces change the conditions of later interaction

An interface can influence the conditions of future interactions. The open question is whether it changes what people attend to, what they consider possible and how they respond to the system, and whether this change feeds back into the data from which a system later estimates states. Interface effects are therefore studied as feedback over time, across several interactions.

Research programs

1 program of its own.

Core phenomena

  • HMI and HCI
  • Relationship between cognition and interface
  • Adaptive interfaces
  • Neuroadaptive direction

Core phenomena

HMI and HCI

  • Human–Machine Interfaces
  • Human–AI Interfaces
  • interaction design
  • interface architecture
  • usability
  • human factors
  • trust through interface design

Relationship between cognition and interface

  • influence on decisions
  • attention
  • mental workload
  • system understanding
  • cognitive scaffolding
  • interface-induced behavioral adaptation
  • cognitive load and interaction patterns
  • human–system feedback loops
  • emergent interaction conventions

Adaptive interfaces

  • context-sensitive interaction
  • state-sensitive interaction
  • adaptive presentation
  • adaptive assistance
  • adaptive system behavior
  • state-dependent interaction

Neuroadaptive direction

  • Behavioral Proxies
  • State Estimation
  • sensor fusion (planned)
  • physiological signals (planned)
  • wearables (planned)
  • EEG (planned)
  • BCI (planned)
  • closed-loop adaptation (planned): human state, system adaptation, feedback, new state
  • technical extension of human capabilities

Keywords (8)

  • interface design and trust
  • neuroadaptive direction (planned)
  • usability and human factors
  • influence on decisions, attention and mental workload
  • context- and state-sensitive interaction
  • EEG and BCI (planned)
  • closed adaptation loop (planned)
  • feedback between human and system

How this space is investigated

Observation and relation

State estimates drawn from behavioral proxies are inferred or hypothesized.

Observation and relation

What is specifically observed

An effect that an interface has on later interaction counts as observed only for the period and conditions in which it was measured. Transfer to other conditions is a hypothesis. Neuroadaptive responses are planned and not implemented.

Further explanation

This space connects classic research on human–machine interfaces with a neuroadaptive research direction.

Relation to the other spaces

This space takes up the question of state-dependent behavior from Human Cognition, Behavior & Externalization, state estimation from Computational Interaction & Cognitive Modeling and the evidence rules from Measurement, Evidence & Research Infrastructure. The program Neuroadaptive Human–AI Systems has its main space here.

Roadmap of the neuroadaptive direction

Five stages lead from interaction data to adaptive system response.

Five stages of the neuroadaptive direction

  1. Interaction dataLongitudinal conversation, turn and session data with timestamps form the only data basis. It exists.available now
  2. Behavioral proxiesBehavioral indicators are derived from the interaction data, such as correction behavior, repetition and strategy shifts. They are indicators. There is no access to mental states.available now
  3. State estimationA state is inferred from behavioral proxies and marked as inferred. Validation against independent measures is pending.experimental
  4. Multimodal sensingPlanned: physiological signals, wearables and EEG/BCI, each only after consent is settled and calibrated against behavioral measures. None of it is collected today.planned
  5. Adaptive system responsePlanned: the system adapts to an estimated state. This requires validated state models and is not implemented at present.planned

On stage 3, State estimation Longitudinal state modeling: estimated states are linked across weeks and years to study transitions and regimes. A concept, not yet built.

The closed loop

The next research frontier is the move from systems that react to explicit user input toward systems that adapt to continuously changing human and environmental states.

The closed loop

Today only behavioral proxies from interaction data are used, neural signals and the adaptive response are planned and not implemented.

Dashed nodes are planned and not implemented. The nodes carry the same three labels as the five stages: available now, experimental, planned.

  1. 01Human cognitive stateplanned
  2. 02Behavioral / neural signalsplannedToday only behavioral proxies from interaction data.
  3. 03State estimationexperimental
  4. 04Human–AI interactionavailable now
  5. 05Agent / intelligence systemavailable now
  6. 06Adaptive responseplanned
  7. 07Human feedbackplanned
  8. 08Updated stateplanned

The updated state closes the loop and leads back to node 01.

Ten preparation rules and five open research questions are with the methods. To the neuroadaptive preparation →

Seven levels of adaptation

  1. 01Interface adaptation
  2. 02Interaction adaptation
  3. 03Workflow adaptation
  4. 04Agent adaptation
  5. 05Memory adaptation
  6. 06Model-selection adaptation
  7. 07Cognitive-state adaptation

Methods, systems and axes

Methods (1)

All methods of this space

Methods and analysis procedures (1)

See all methods →

Systems (2)

All systems of this space

Research systems (1)

Shared technical foundation (1)

See all systems →

Cross-cutting research axes (1)

How the axes act in this space

Cross-cutting condition Ethics / Human Agency / Human Rights →

Evidence & limits

Evidence boundary

Planned neuroadaptivity may be described clearly as a research direction but never as an existing measurement capability.

Available today
  • interaction data
  • behavioral proxies
Planned, not implemented
  • physiological signals
  • wearables
  • EEG
  • BCI
  • full adaptive response

A state estimate is not a measurement of the state

A behavior pattern assigned to a state is an inferred estimate and not a measurement of that state. Uncertainty is not treated as a certain state. Interaction conventions count as observed only once they recur and persist across contexts.

Boundary of this space

Measurable signals and interface responses do not allow an unambiguous inference of complex mental states. Adaptive systems must not treat uncertainty as a certain state.

How to read status and origin

Every statement carries its status (observed, inferred, hypothesis, planned or unknown), no program is externally validated, and origin statements about systems are candidates.

Evidence register → Origin of the systems →

Topics and publications

working formulation Formulated from the research program, not yet a fixed wording.

Topics with this space as main space (2)

Further topics that touch this space (1)

Guiding questions of the topics

  • 04 Neuroadaptive Human–AI Systems

    How can systems adapt to changing cognitive states while keeping measured and inferred states apart?

  • 12 Human–Machine Interfaces

    How do interfaces shape cognition, trust and system understanding, and what will an interface need to show?

  • 11 Adaptive Systems

    How does adaptive behavior stay inspectable and traceable?

Publications (1)

Publications (1)