Author: artun.sepken

SIGCHI Austria Fall Meeting | 26 September 2026 | University of Vienna

As I continue my master’s studies in Interaction Design, I have been trying to understand where I want to take my research and what opportunities exist beyond my studies. My background is mainly in visual communication and design, but over the past year, I have become increasingly interested in Human–Computer Interaction (HCI), particularly in the research side of the field. This interest led me to join SIGCHI Austria and attend its Fall Meeting, which took place alongside uxcon vienna at the University of Vienna.

The meeting brought together researchers, PhD students, postdoctoral researchers, professors, and people working in different areas of HCI. The programme included research presentations, keynote talks, discussions, and opportunities to meet other participants. It was a full day of presentations and conversations, but for me, the most valuable part was getting to know the community and seeing how people with different backgrounds approach interaction research.

A different perspective on interaction design

One thing I noticed quite quickly was that my background was different from those of many other participants. Most of the people I spoke with were involved in academic research, often coming from technical or scientific disciplines. I was approaching HCI mainly through design, with an interest in how interaction concepts can be developed, prototyped, and evaluated.

This made the experience particularly interesting. I had several conversations about my studies, my interests, and the kind of research I might pursue in the future. Speaking with people who were already doing PhDs, working as postdoctoral researchers, or teaching at universities helped me understand the different paths within HCI.

I also began thinking more about the relationship between design and research. In design education, we often work towards an outcome, whether it is an interface, an experience, or an interactive prototype. In academic research, the contribution can also be a better understanding of a problem, an experimental finding, or a method that others can build upon.

These are not completely separate approaches, but I am still learning how to bring them together in my own work.

Research presentations and discussions

Throughout the day, I attended the research presentations and discussions organized by SIGCHI Austria.

What I appreciated about the format was that each presentation was followed by discussion. The audience could ask questions, share suggestions, and discuss possible directions for the research. It was interesting to see research being discussed while it was still developing, rather than only encountering it as a finished publication.

Another highlight was Elizabeth Churchill’s keynote, “Intelligence Reimagine – Interaction Redefined,” presented as part of uxcon vienna. Churchill has worked across academic HCI research and industry, including leading UX research teams at Google and other technology companies. Her background alone was interesting to me because it demonstrates how research and design practice can be connected throughout a career.

I particularly enjoyed her presence and the way she communicated. She was an engaging speaker with a strong personality, which made her presentation one of the most memorable parts of the event. Her keynote addressed the changing relationship between people, technology, and design in the context of AI. These questions are also relevant to how I am currently thinking about interactive systems.

Finding my place as a Interaction Designer

Beyond the presentations, I think the conversations I had throughout the day were just as important. Meeting people at different stages of their academic careers made the research community feel more accessible.

Before attending, I was familiar with HCI mainly through papers-projects, and the research methods I had been studying. At the meeting, I could see the people behind this work, hear about different research directions, and ask questions about their experiences.

I also offered to contribute to SIGCHI Austria through volunteering, particularly with design, organization, and related activities. I would like to stay involved because I see it as a way to learn from the community while contributing something from my own background.

The experience also made me more confident about approaching researchers, even when I do not have the same technical or academic background. I still have a lot to learn about research methodology and academic work, but I can bring my own perspective as a designer into these conversations.

Impact on my further research

This meeting came at an important point in my studies, as I have been exploring possible directions for my master’s thesis and considering further academic research.

Attending the meeting helped me see the broader field in which my interests could develop. It reminded me that HCI is interdisciplinary and that meaningful contributions can come from different methodological and professional backgrounds.

Perhaps the biggest takeaway was realizing that I do not have to choose between being a designer and becoming a researcher. I can develop both perspectives, and the combination may become an important part of my future work.

Attending SIGCHI Austria was therefore more than an opportunity to listen to presentations. It was my first direct experience of becoming involved in an academic HCI community, and it encouraged me to continue exploring this direction.

---

Accompanying links

1. SIGCHI Austria — Autumn Meetup 2026 Official information about the meeting and its collaboration with uxcon vienna.
2. uxcon vienna 2026 — Day 1 Programme Conference programme, including Elizabeth Churchill’s keynote.
3. Elizabeth Churchill — Speaker Profile Information about her academic research, professional background, and keynote.

AI Use Disclaimer

AI was used to help organize and edit this reflection. The personal experiences, observations, and reflections are my own.

For this assignment, I chose Adaptive Affective Computing: Countering User Frustration by Behzad Aghaei (2013), written at the University of Waterloo. I chose this thesis because it relates to my interests in human-computer interaction, cognitive workload, and adaptive interfaces. I am particularly interested in how interactive systems can recognize changes in a user’s state and respond to them in a useful way.

About the Thesis

The thesis explores whether an adaptive interface can help users recover from frustration. The author tested this using a tablet-based map application that deliberately froze during tasks. Participants wore a skin conductance sensor to measure changes in their physiological arousal. In the adaptive condition, the interface changed its colors and button sizes as part of its response to frustration.

The thesis follows a clear structure, starting with the theoretical background and continuing with the experimental methodology, data analysis, results, and discussion. The practical part involves a working prototype and an experiment comparing a standard interface with an adaptive version. The main idea is to investigate whether changing the interface based on the user’s physiological state can improve their experience.

Academic Quality and Practical Implementation

I think the research question is interesting, especially considering that the thesis was written in 2013. The idea of using physiological signals to influence interface behavior was an experimental approach at the time. I also appreciate that the project goes beyond measuring frustration and attempts to respond to it through interaction design. Although the adaptation itself is relatively simple, the research raises an important question about whether interfaces should automatically react to users’ emotional states.

The experimental approach is one of the strengths of the thesis. The author compared two interface conditions and collected both physiological measurements and participant feedback. This provides different perspectives on the user experience and helps connect the theoretical ideas to measurable results. However, there are some limitations in how the experiment was conducted.

The study initially included 40 participants, but only 22 were included in the final analysis. Although the author explains the reasons for excluding participants, this is still a significant reduction in the sample size. It raises questions about how reliable the findings are and whether similar results would appear with a larger group. Another important issue is that the frustration detection algorithm rarely worked as intended. I think this is particularly problematic because reliable detection is an essential part of the proposed adaptive system. If the system cannot consistently recognize the user’s state, it becomes difficult to evaluate the full potential of the adaptation.

The practical project is documented through screenshots, explanations of the experimental procedure, and graphs. The comparison between the standard and adaptive interfaces is especially helpful for understanding the implementation. Visually, the interface looks functional, but I do not find the design particularly convincing. The adaptation mainly involves color and button size changes, without much exploration of how these decisions affect the overall interaction. The visual changes also seem quite noticeable, which could distract users rather than support them. I think testing different adaptation strategies and conducting usability tests before the main experiment could have made the practical work stronger.

In terms of organization and communication, the thesis is generally clear and easy to follow. The transition from theory to experiment and results is logical, and the author explains the technical concepts in a way that supports the research. However, some parts of the theoretical background feel longer than necessary. I think a more focused discussion could have made the connection between the literature and the actual design decisions clearer.

The scope is also quite ambitious, combining psychology, physiological sensing, software development, and experimental evaluation. This makes the project interesting, but it also creates technical and methodological challenges. In my opinion, focusing more on the quality of the adaptation and its evaluation might have been more valuable than trying to cover so many areas at once.

The literature covers relevant academic fields and provides a useful foundation for the research. The author uses previous studies to explain physiological responses, frustration, and adaptive computing. However, I think the assumptions behind the visual adaptation could have been examined more critically. For example, there is a difference between knowing that certain colors may have calming associations and proving that changing an interface to those colors will actually help someone recover from frustration.

The thesis is generally presented carefully, with references, figures, and methodological explanations. The bibliography is relevant to the research topic, although the connection between some of the literature and the final design decisions could be stronger. I did not notice major problems with the overall presentation, but some methodological decisions could have been justified more clearly, particularly the participant exclusions.

Regarding independence, the author acknowledges that a colleague developed the software system. I do not see collaboration as a problem, especially in a technical project, but a clearer explanation of the author’s individual contributions to the practical work would have been helpful. The research direction, experimental approach, and analysis still demonstrate an independent academic investigation.

Results and Overall Evaluation

What I found most interesting is that the adaptive interface did not improve recovery as expected. The overall difference between the two groups was not statistically significant, and the first trial actually showed slower recovery in the adaptive condition.

Some participants also associated the visual changes with the application freezing rather than understanding them as a supportive response. I think this is an important finding from an interaction design perspective. Even when a system is designed to help users, the way it communicates its behavior can create confusion. The intention behind an interaction does not always match how people experience it.

Overall, I think the thesis meets the expectations of a master’s-level research project. It connects a research question with theoretical knowledge, practical implementation, and experimental evaluation. Its main weaknesses are the reduced sample size, unreliable frustration detection, and limited exploration of alternative designs.

However, I appreciate that the author discusses the unexpected results and possible reasons behind them. The prototype did not achieve its intended outcome, but the experiment still provides useful insights into the challenges of adaptive interaction. I think the academic value of the thesis comes more from its research process and findings than from the success of the prototype itself.

Relevance to My Own Thesis

This thesis connects closely to my interests in adaptive interaction, attention, and cognitive workload. I am interested in systems that can respond to users’ changing needs without interrupting their activities unnecessarily.

One thing I would like to take from this thesis is the use of a controlled experiment with clear hypotheses and measurable results. I also find the combination of physiological data and user feedback useful because it allows researchers to compare what a system measures with what users actually experience.

However, I would approach the design process differently. Before conducting the main experiment, I would spend more time testing how users understand and react to the adaptation itself. I would also consider combining physiological sensing with behavioral measurements instead of relying on a single signal. This could provide a more complete understanding of the user’s state and help avoid interpreting physiological arousal as one specific emotion.

Another aspect I would pay attention to is how noticeable an adaptation should be. In some situations, changing the interface might help users, but in others, the change itself could become an interruption. I think this is especially important when designing systems intended to support attention and task continuity.

The most important lesson for me is that an adaptive interface is not automatically a better interface. A system might successfully recognize a change in the user’s state, but its response can still be confusing or distracting. For my own research, I would like to focus not only on whether a system can adapt, but also on whether that adaptation actually supports the user’s experience.

Reference

Aghaei, B. (2013). Adaptive affective computing: Countering user frustration [Master’s thesis, University of Waterloo]. https://hdl.handle.net/10012/7386

Google Drive Link

In the last post I argued that the research had to stop being a broad map and become one concrete thing I could build and test. This is that thing.

I built a neuroadaptive supervision console. On screen, the operator keeps a small fleet of semi-autonomous drones alive: each one occasionally needs attention and timed emergencies appear that must be resolved before a countdown runs out. The drones are not the point for me, they are a stand-in for any complex, information-dense interface, the kind of control-room or monitoring setting where alerts compete for attention and overload is a constant risk. That is where badly-timed interruptions do the most damage, so that is the situation I wanted to recreate.

Underneath the task, the system builds a live estimate of how loaded the operator is, by combining three real signals: EEG engagement from a Muse headband, a webcam-based attention proxy (for blink rate and head movement – unfortunately I could not extract data from Tobii) and the demand of the task itself. I then compared two versions of the same interface. In the adaptive version, when the estimate crosses into “overload,” the system quietly defers low-urgency alerts until there is room for them and steps in for the drones most at risk. In the control version it does none of that. Every participant played both.

The clearest result, shown in the video, is reaction time. On the working build, people answered emergencies far faster in the adaptive version – around four seconds, against seventeen in the control. The mechanism is simple: when alerts are not deferred, they pile up during busy moments and get answered at the last second. Deferring them spreads the load out.

What the video does not point out on is that this was not a clean win and I think that matters more than the headline. Faster answers came with lower accuracy, and people did not report feeling less loaded. With only four participants, and a build that changed between the early and late sessions, this is a proof-of-concept, not proof. The fuller per-person picture, including a resumption-lag result that went the “wrong” way, sits in my written results rather than the video.

For me the takeaway is not the number but the shape of the problem. A real adaptive loop is buildable with modest, honest signals, and once it works the interesting questions are not about better sensors. They are about timing, trust and giving the person a way to overrule the system when its guess is wrong. That is the thread I want to keep pulling: from interfaces that capture attention toward interfaces that look after it.

After the UX Graz talk, I was left with a useful tension. The research had developed into a broad field of connected questions, but the next step needed to be more concrete. My professor encouraged me to experiment rather than force the work into a final thesis topic too early. This semester is a chance to build something, test one direction and learn from what does not work.

What I want to experiment with is neuroadaptive interaction.

This does not mean that neuroadaptive interfaces have become my final thesis direction. Adaptive interaction design is still the larger area I am exploring. My interests include interruption, recovery, cognitive workload and the changing relationship between people and increasingly proactive systems. Neuroadaptive interaction gives me one way to bring these interests together in a form that can be experienced rather than only described.

I have been especially drawn to experimental HCI research from the 1990s and early 2000s. Before current discussions about AI agents and intelligent assistants, researchers were already asking how systems might respond to changing human capacity. Work on mixed-initiative interfaces, adaptive automation and physiological computing treated interaction as a continuing feedback loop. The system observed something about the user or task, adjusted its behaviour and then observed the consequences.

Pope, Bogart and Bartolome’s (1995) biocybernetic research is a good example. Their system used an EEG-derived engagement index to change the balance between manual and automated work in a flight-deck task. The experiment belongs to a specialised context and its measurements should not be treated as a direct model for everyday interfaces. Still, the interaction question remains valuable. If a system changes its level of support according to an estimate of the user, how should that change be communicated and how can the user remain in control?

Horvitz’s (1999) work on mixed-initiative interfaces approaches a similar problem without making physiological sensing the centre. A system may be able to suggest an action or offer help, but that does not mean every moment is suitable for intervention. It has to decide when to act, when to ask and when to remain quiet. These choices distribute initiative between the person and the system.

The historical work interests me because it is experimental in a direct way. It does not only propose that interfaces could become more responsive. It constructs situations where the relationship between user state, task demand and system behaviour can be observed. This is the kind of step I want to take now.

The experiment is not about proving that a computer can read somebody’s mind. Neuroadaptive systems work with indicators and inferences. EEG, gaze, behaviour and task performance can suggest that something may be changing, but they do not provide uncomplicated access to attention or mental workload. Fairclough (2009) describes a biocybernetic loop as a process of sensing, inference, adaptation and feedback. The inference stage is where uncertainty enters. One physiological change can have several possible causes, and the same cognitive state can appear differently across people.

Treacy Solovey et al. (2015) make the design gap explicit when they write that “the interaction techniques and design decisions for their effective use are not well defined” (p. 1). Their work suggests that physiological signals may be more appropriate as implicit context than as direct commands. Adaptations should be modest, confidence should matter and changes should remain reversible.

This shifts my attention from the accuracy of a single sensor toward the behaviour of the interface. What should a system do when it estimates that workload may be rising? Should it reduce information, delay a low-priority interruption or increase automation? How much should it explain? What should happen when its estimate is wrong?

To examine these questions, I need a setting where task demand and adaptation can both become visible. A simple interface would not create enough pressure to make the experiment meaningful. If a person is only watching one value or completing one predictable action, there is little reason for the system to adapt. The context needs several competing demands without becoming impossible to understand.

This is why I became interested in complex interfaces as an experimental environment. Control rooms, monitoring systems and operational dashboards require people to maintain awareness across several sources of information. Some events are urgent while others can wait. The interface may need to guide attention without hiding information or taking control away from the operator.

Complex interfaces also make cognitive load a design issue rather than an abstract score. Information density, competing priorities and time pressure can all affect what the user is able to process. An adaptive system could respond by foregrounding the most relevant event, reducing noncritical detail or temporarily taking over a routine action. Each of these adaptations can help, but each can also create new problems. Removing information may reduce visual load while weakening situation awareness. More automation may lower immediate demand while making it harder for the user to understand what the system has done.

The experiment therefore needs to be about control negotiation, not automatic optimisation. The user should be able to inspect why the system changed, reject the change or disable adaptation. An uncertain estimate should never give the system unquestionable authority. If the prototype adapts, it should do so in a way that keeps its reasoning and limitations visible.

Interruption and recovery remain part of this experiment. In a complex interface, the system has to decide whether a new event deserves immediate attention. Adamczyk and Bailey (2004) showed that interruption timing affects performance and emotional response. Chen et al. (2025) demonstrate that recovery cues can support users after an interruption by directing them toward previous or upcoming task information. These findings suggest several interactions that can be tested without pretending to solve cognitive workload as a whole.

The first version can begin with simulated signals. This allows me to test what the interface does before depending on hardware. Later, a Tobii eye tracker and a Muse headset may provide live inputs, but they should not define the project. A real sensor cannot rescue an adaptation that is confusing, intrusive or impossible to challenge.

For a concrete implementation, I am using a drone-triage scenario. The user monitors several semi-autonomous drones while managing interruptions and changing priorities. This setting provides a complex interface where workload, timing, assistance and recovery can be represented clearly. The system can defer a lower-priority alert, foreground a risk or provide a cue after the operator returns to the main task.

The drone scenario is not the research decision. It is a practical container for the experiment. Another operational context might eventually work better. What matters is that the interface gives me a way to test neuroadaptive behaviour under visible task pressure while preserving explanation and human override.The aim is to make one part of the research concrete enough to question, observe and revise.

References

Adamczyk, P. D., & Bailey, B. P. (2004). If not now, when?: The effects of interruption at different moments within task execution. In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems (pp. 271-278). ACM. https://doi.org/10.1145/985692.985727

Chen, Y., Zhang, C., Fang, W., & Ma, J. (2025). The effects of cues on task interruption recovery in a concurrent multitasking environment. Scientific Reports, 15, 25992. https://doi.org/10.1038/s41598-025-09358-4

Fairclough, S. H. (2009). Fundamentals of physiological computing. Interacting with Computers, 21(1-2), 133-145. https://doi.org/10.1016/j.intcom.2008.10.011

Horvitz, E. (1999). Principles of mixed-initiative user interfaces. In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems (pp. 159-166). ACM. https://doi.org/10.1145/302979.303030

Pope, A. T., Bogart, E. H., & Bartolome, D. S. (1995). Biocybernetic system evaluates indices of operator engagement in automated task. Biological Psychology, 40(1-2), 187-195. https://doi.org/10.1016/0301-0511(95)05116-3

Treacy Solovey, E., Afergan, D., Peck, E. M., Hincks, S. W., & Jacob, R. J. K. (2015). Designing implicit interfaces for physiological computing: Guidelines and lessons learned using fNIRS. ACM Transactions on Computer-Human Interaction, 21(6), Article 35, 1-27. https://doi.org/10.1145/2687926

During this research process, I was invited to give a talk at UX Graz. The invitation came at an interesting point in my work. I had spent months reading, I had a lot of material, many connections and a growing sense about the topic. At the same time, I was still exploring rather than defending one narrow thesis topic.

The talk became an opportunity to test whether the research could make sense outside my own notes and Zotero library.

I titled it Designing for the Interrupted: Notes on Attention and Adaptive Interfaces. I did not want to present it as a finished argument or a final solution. The aim was to share the direction of the research with a UX audience and to ask what we lose when digital systems treat every moment as equally interruptible.

The central thought behind the talk was simple: interruptions are not small interface events. They affect task continuity, emotional state, workload and the ability to return to something afterwards. A notification may only take a few seconds to read, but the cognitive work of resuming a task can last much longer. This is why the design question cannot only be whether an alert is visually clear or technically delivered. It also has to be whether it arrived at a moment when the user could reasonably deal with it.

To build this argument, I began with attention and flow. Flow is often used loosely in design conversations, sometimes as another word for engagement. What interested me in Csikszentmihalyi’s work was the structure behind it: clear goals, feedback, a relationship between challenge and skill and a sense of control over the activity. These conditions are easy to disrupt. An interface that repeatedly changes the user’s context can make concentration feel fragile even when every individual interruption appears minor.

From there, the talk moved through a series of research areas that had shaped my work. Interruption studies show that timing changes the cost of disruption. Cognitive-load research shows that people are not equally interruptible in every situation. Emotional-design research shows that interruptions can communicate pressure, urgency and social obligation rather than merely information. Memory and resumption research show that the real problem often begins after the interruption, when people have to reconstruct what they were doing.

I also wanted to avoid the common idea that this is a “short attention span” problem. Attention is not a fixed amount of time that every person has before they stop listening or working. It is contextual. It changes with task structure, motivation, environment, cognitive control and the behaviour of the system itself. The problem is not that users have become incapable of focusing. The problem is that many interfaces are built to claim attention whenever they can.

Notifications were the most recognisable example in the talk, but I tried not to make them the whole subject. They are only one visible form of interruption. A system can interrupt through sound, vibration, visual prompts, a change in automation, a request for approval or a recommendation that appears at the wrong time. This becomes even more important as interaction moves beyond a laptop screen into wearables, mixed reality and AI-agent systems.

The talk also gave me a way to discuss neuroadaptive interfaces without treating them as magic. Physiological computing and workload-aware interfaces offer an interesting interaction logic: a system can use signals from the body, behaviour or task performance to adjust its own behaviour. But this only makes sense if the system acknowledges uncertainty. It should not pretend that gaze data or EEG can reveal a person’s complete mental state. It should use partial signals carefully, explain what it has changed and let the user override it.

The final part of the talk connected this to current AI systems. AI agents create a new version of an old design problem. If an agent constantly asks the user for confirmation, it can become another source of interruption. If it acts independently for too long, it can create errors that are difficult to discover and correct. The human role cannot be reduced to pressing approve or reject. It has to include understanding what the system is doing and deciding when to step in.

Preparing the talk was valuable because it forced me to see the research as one larger landscape. It also made the limits of that landscape visible. The live presentation was research-rich, but it covered too much for a twenty-minute setting. The feedback afterwards was fair and consistent. People found the topic interesting, but they needed one concrete example or object to hold onto. The talk felt like several connected research directions compressed into one argument.

I do not see that feedback as a failure. It was a scope diagnosis. The research was not empty, but it was still too broad to communicate through one presentation alone. That distinction matters. A broad research map can be useful in the early stages of a project, yet it eventually has to lead to a specific situation where its ideas can be tested.

The strongest phrase I used near the end of the talk was “from capturing attention to caring for it.” I still believe that this is the right design direction. It does not mean that interfaces should never interrupt, automate or ask for something. It means they should recognise that attention has a cost and that the user should remain part of the decision about when that cost is worth paying.

The UX Graz talk gave me a public moment to articulate that position. More importantly, it gave me a reason to move forward differently. Instead of trying to make the next step another broad explanation of the field, I want to experiment with one concrete adaptive interaction loop. That is where the prototype begins.

Also linking the presentation here in case someone is interested: https://www.figma.com/deck/knB8wTii1xOntnhMrQMYgN

References

Adamczyk, P. D., & Bailey, B. P. (2004). If not now, when?: The effects of interruption at different moments within task execution. In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems (pp. 271-278). ACM. https://doi.org/10.1145/985692.985727

Altmann, E. M., & Trafton, J. G. (2002). Memory for goals: An activation-based model. Cognitive Science, 26(1), 39-83. https://doi.org/10.1207/S15516709COG2601_2

Csikszentmihalyi, M. (1990). Flow: The psychology of optimal experience. Harper & Row.

Fairclough, S. H. (2009). Fundamentals of physiological computing. Interacting with Computers, 21(1-2), 133-145. https://doi.org/10.1016/j.intcom.2008.10.011

Iqbal, S. T., & Horvitz, E. (2007). Disruption and recovery of computing tasks: Field study, analysis, and directions. In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems (pp. 677-686). ACM. https://doi.org/10.1145/1240624.1240730

Lavie, N. (2010). Attention, distraction, and cognitive control under load. Current Directions in Psychological Science, 19(3), 143-148. https://doi.org/10.1177/0963721410370295

In the previous post, I arrived at a direction that feels more specific than the topic cluster I carried through last semester: adaptive interaction design. I am not treating it as a finished thesis definition. It is a working direction that gives my reading and prototyping somewhere to go.

That shift changes how I look at the literature. Before, I was reading broadly in order to understand interruption, flow, cognitive load, memory, emotion and neuroadaptive interfaces. The breadth was necessary. It helped me see that interruptions are not isolated design failures but events that affect attention, task continuity and a person’s ability to return to what they were doing. Now I need to ask a different question. If an interface is going to adapt to the user, what should it pay attention to, what should it be allowed to change and how should the user remain involved?

The older work on adaptive and mixed-initiative systems has become especially useful here. It is easy to speak about adaptive interfaces as if they are a recent consequence of AI, but many of the central interaction questions are older. Horvitz (1999) wrote about mixed-initiative user interfaces at a time when the examples looked very different from current AI agents. Yet the core problem is familiar: a system may have enough information to make a suggestion or take an action, but it still has to decide whether this is the right moment to involve the user.

That matters because a useful system can also become a badly timed system. It may offer a recommendation while somebody is already working through a difficult task. It may ask for confirmation when the person has no real capacity to evaluate the decision. It may quietly act on the person’s behalf, then leave them to reconstruct what happened later. The design problem is not simply how to make the system more proactive. It is how to distribute initiative without turning the user into a passive observer of their own work.

This connects directly to the research I had already done on interruption. Adamczyk and Bailey (2004) showed that the timing of an interruption changes its effects on performance, emotion and the way people perceive the interrupting system. Their basic point remains important: “user attention is a scarce resource” (p. 271). An interface cannot assume that a message, recommendation or request for approval has the same cost at every point in a task.

Iqbal and Horvitz (2007) make the picture more realistic by looking at interruption and recovery in everyday computing work. Their work shows that disruption is not just the moment somebody looks away from a task. There is a longer process of reorientation afterwards. People need to recover their goal, remember the current state of the work and decide what comes next. This is why I am interested in adaptation not only as a way to prevent badly timed interruptions, but also as a way to support the return after they happen.

The return is where memory becomes a design concern. Altmann and Trafton (2002) describe pending goals in terms of activation and retrieval. In simple terms, an interrupted goal does not wait untouched until a person comes back to it. It has to compete with everything else that happened in the meantime. The task may still be open on the screen, but the reasoning behind it may have faded. A system that changes information, makes suggestions or completes work while the user is away has a responsibility to make the return understandable.

This becomes particularly relevant in AI-assisted work. Recent research on agentic systems is often framed around how much work an agent can perform autonomously. I find the moment of human re-entry more interesting. Zhou et al. (2026) examine when people should check multi-step agentic tasks. Their findings suggest that constant confirmation becomes costly, while waiting until the end can make correction much more expensive. The issue is not whether the human should supervise every action. It is how the system can involve the human at moments that protect both control and continuity.

The same question appears in workload-aware systems. The literature does not support a simple idea of a system knowing exactly when someone is busy. Workload overlaps with attention, task difficulty, stress, effort and fatigue, but it is not interchangeable with any of them. Kosch et al. (2023) show that HCI researchers use self-report, performance, behaviour and physiological measures, each with different strengths and limits. There is no single number that can tell an interface everything it needs to know about a person.

That uncertainty is not a reason to abandon adaptive design. It is a reason to design it carefully. Fairclough’s (2009) work on physiological computing describes a loop in which a system senses something about the user, infers a state, adapts and then observes the result. The word infers is important. Signals from gaze, EEG, typing or task performance are not direct access to a person’s internal life. They are partial evidence that may or may not be useful in context.

I find the interaction logic of this early experimental work more compelling than the promise of any individual sensor. The system does not need to claim that it has discovered a person’s true mental state. It can respond modestly to patterns that suggest a change in task pressure, then let the user inspect or reject that response. Treacy Solovey et al. (2015) argue that physiological input is often better suited to implicit and supportive adaptation than direct control. Their point is not that systems should become invisible. It is that they should not behave with more certainty than their input allows.

This is starting to give my research a clearer shape. I am not looking for the perfect way to measure attention. I am looking at what happens when an interface has incomplete information about the user’s situation and still needs to make decisions about timing, assistance and interruption. What would it mean for that interface to be helpful without becoming intrusive? How could it support recovery without taking ownership away from the person?

The answers will depend on the context. A work tool, a wearable device, a spatial interface and an AI agent all create different conditions. For now, I do not want to lock the research into one medium too early. I want to keep reading across these contexts while holding onto the same underlying question: how can systems adapt to human capacity without asking humans to adapt to the system first?

This post is a research checkpoint rather than a conclusion. The previous post helped me name a direction. This one is helping me understand the problems inside that direction.

References

Adamczyk, P. D., & Bailey, B. P. (2004). If not now, when?: The effects of interruption at different moments within task execution. In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems (pp. 271-278). ACM. https://doi.org/10.1145/985692.985727

Altmann, E. M., & Trafton, J. G. (2002). Memory for goals: An activation-based model. Cognitive Science, 26(1), 39-83. https://doi.org/10.1207/S15516709COG2601_2

Fairclough, S. H. (2009). Fundamentals of physiological computing. Interacting with Computers, 21(1-2), 133-145. https://doi.org/10.1016/j.intcom.2008.10.011

Horvitz, E. (1999). Principles of mixed-initiative user interfaces. In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems (pp. 159-166). ACM. https://doi.org/10.1145/302979.303030

Iqbal, S. T., & Horvitz, E. (2007). Disruption and recovery of computing tasks: Field study, analysis, and directions. In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems (pp. 677-686). ACM. https://doi.org/10.1145/1240624.1240730

Kosch, T., Karolus, J., Zagermann, J., Reiterer, H., & Schmidt, A. (2023). A survey on measuring cognitive workload in human-computer interaction. ACM Computing Surveys, 55(13s), Article 286, 1-39. https://doi.org/10.1145/3582272

Treacy Solovey, E., Afergan, D., Peck, E. M., Hincks, S. W., & Jacob, R. J. K. (2015). Designing implicit interfaces for physiological computing: Guidelines and lessons learned using fNIRS. ACM Transactions on Computer-Human Interaction, 21(6), Article 35, 1-27. https://doi.org/10.1145/2687926

Zhou, Y., et al. (2026). When should users check? Modeling confirmation frequency in multi-step agentic AI tasks. In Proceedings of the CHI Conference on Human Factors in Computing Systems. ACM.

In my first blog post of this semester, written after the lo-fi prototyping session, I described a frustration that felt more important than the prototype itself. Moving from theoretical research into a quick physical test forced me to confront something I had been postponing: my topic is difficult to make visible. Attention, interruption and cognitive recovery are not things you can easily point at on a screen. They happen internally, across time and often become visible only through their consequences.

At the time, that post was mostly about the limitations of the format I had tested. Looking back, it was also the beginning of a more important question. If the topic I am researching is hard to prototype, then maybe the issue is not only the prototype. Maybe I had not yet defined clearly enough what kind of design problem I was actually trying to make visible.

This post continues that questioning.

Where the First Semester Left Off

Last semester I wrote ten posts about attention, flow, interruption, memory, emotion, and neuroadaptive interfaces. Taken one by one they look scattered. Taken together, they were building toward something I could not yet name.

The clearest position I reached was in my final post: interruption is not a problem to eliminate, it is a condition to design for. What matters is not whether a system interrupts, but whether it does so at the right moment, with the right cost, and with enough support for the user to recover.

That framing felt like a conclusion at the time. Looking back, it was more like a door. The research had built a detailed picture of what goes wrong attention fractures, cognitive load spikes, emotional friction accumulates, memory decays but it had not yet said what to do about it, in what context, using what kind of design.

That is what this semester is supposed to resolve.

A Detour That Wasn’t Really a Detour

Earlier this semester I joined a project connected to CERN’s IdeaSquare initiative through our university. My group was working on individual and communal spaces, specifically the idea of modular, adaptive environments that can shift and reconfigure themselves in response to the people using them.

At some point during the project I pushed the team toward a particular angle: spaces that adapt to their users, not the other way around. The environment should read who is there, what they are doing, and what they need, then reorganize accordingly. Less about static design and more about responsive systems.

I did not connect this to my research at the time. It just felt like the right direction for that project.

Later I realized I had done exactly the same thing I had been studying. The question I brought into that spatial design project was the same question underneath my first semester work: what would it look like if systems adapted to human state, rather than expecting humans to adapt to them? In the CERN project, the system was a physical space. In my research, it had been a digital interface. But the logic was identical.

That overlap was worth noting. Not because it proves anything, but because it confirmed something: this is not an academic interest I adopted for a class. It is genuinely how I think about design.

The Problem With Where I Was

After last semester, I had a topic cluster but not a research direction. There is a real difference. A topic cluster is a set of connected ideas you can read about indefinitely. A research direction has a specific question, a context, and a method.

My cluster was attention, interruption, cognitive load, recovery, neuroadaptive systems. Interesting and defensible, but I was circling the same ideas without committing to a specific design problem.

The first blog post of this semester already hinted at this. The lo-fi prototype did not fail because the topic was weak. It struggled because I was trying to represent an internal and temporal topic before fully deciding what form of intervention I was actually designing toward. The session was useful precisely because it exposed that gap.

The second semester also added a different kind of pressure. The business-framing exercises pushed me to specify who is affected, how and what a system would actually do. Together with the prototyping difficulties, this made clear that narrowing was not optional anymore.

What I Actually Want to Research

The direction I have arrived at is best described as adaptive interaction design.

Not notification management. Not UX for focus apps. The broader question is what it would mean for interfaces to adapt their behavior based on the user’s current cognitive state, rather than treating every moment as equally available for input, output, or interruption.

The interest in neuroadaptive and EEG-based systems from last semester was never really about brainwaves. It was about the interaction logic they represent: a closed-loop system that reads the user continuously and responds accordingly. That loop, not the hardware, is what I find compelling. The philosophical stance is that machines should adapt to humans, not the other way around. The sensor stack is just one implementation of that idea.

What has become clearer this semester is that newer forms of adaptive systems make this question even more urgent. When systems act, update, or reorganize information while the user is absent, the moment of return becomes a design problem in itself. My earlier research on interruption, memory, and resumption maps directly onto this. The cognitive science is already there. What is still missing is a clear design response.

Where This Is Heading

I am not committing to a final thesis question this semester. That is not what this phase is for. What I am committing to is a direction: adaptive interaction design, specifically the intersection of cognitive state, interruption, and resumption in contexts where systems change while the user is absent.

The medium is still open. It could remain screen-based. It could later move into spatial or wearable contexts if the research actually needs that. I do not want to decide the form too early.

What I want to do this semester is go deeper into the literature on workload-aware systems, neuroadaptive interaction, and newer adaptive interface models, then develop a prototype concept that makes the re-entry problem testable. Not to solve it immediately, but to make it visible enough to study.

The first blog post of this semester started with a prototype that revealed the limits of my current framing. This post is the continuation of that realization. The question is no longer only how to prototype interruption and recovery. It is how to define a research direction where adaptation itself becomes the design material.

That is the point I seem to have reached. Now the next task is to see whether I can narrow it without losing what made it interesting in the first place.

Customer Profile & Value Proposition Map

Persona I – The Deep Worker

A knowledge worker or student who needs extended periods of uninterrupted focus to do their best work. They are productive in flow states but regularly pulled out of them by notifications, messages, and contextual switches. They need a system that protects their attention without requiring constant manual management.

Persona II – The Overloaded Student

A university student juggling coursework, communication apps, and social media across the same device. They struggle to distinguish between urgent and non-urgent signals, and often spend more time managing notifications than doing the actual work. They need a system that reduces the noise without them having to think about it.

Business Idea

What problem are you solving?

Digital interfaces are built to deliver information as fast as possible: but human attention does not work that way. Every interruption carries a cognitive recovery cost that current systems completely ignore. The result is a generation of users who are constantly reactive, chronically distracted, and unable to reach the deep focus states where their best thinking happens. There is no mainstream product that treats attention as a resource worth protecting at the system level.

Why should we care about it?

Attention is not just a productivity concern, it is a mental health issue. Sustained notification overload is linked to higher cortisol levels, reduced working memory performance, and increased anxiety. At the same time, the economic cost of fragmented attention in knowledge work is measurable: studies estimate billions in lost productive hours annually due to interruption-driven task switching. The problem affects every person who works or studies with a digital device which is effectively everyone.

What is the solution? How does it work?

An attention-aware notification layer that sits between the operating system and the user’s apps. It uses behavioral signals: typing rhythm, app dwell time, task duration, time of day, to infer whether the user is in a focused state. When focus is detected, non-urgent notifications are held and batched for delivery at a natural task boundary. When an interruption does occur, the system provides a resumption cue, a lightweight context snapshot that helps the user return to their previous task faster. No manual configuration required; the system learns the user’s patterns over time.

Who is the target audience / customer?

The primary users are knowledge workers and students: anyone whose productivity depends on sustained focus. The paying customers are organizations: companies that want to reduce burnout and increase deep work capacity among employees, educational institutions looking to support student focus, and productivity software companies that want to integrate attention-awareness into their existing tools as a premium feature.

What is going to happen? (Change & Impact)

We move from a model where every moment is equally interruptible to one where digital systems respect the rhythm of human cognition. Interruptions do not disappear, they are timed better. Users reclaim extended focus periods without having to fight their devices to do it. Over time, this reduces the normalization of fragmented attention and establishes a new expectation: that technology should protect focus, not just compete for it.

System Map

At the center is the Attention-Aware Interface System, a framework designed to make digital environments sensitive to a user’s attentional state before delivering interruptions, rather than eliminating them entirely. The inner ring holds the direct users: knowledge workers, students, remote workers, and multitaskers, each experiencing interruptions differently based on task type and cognitive load. Surrounding them are the groups that shape interruption conditions: app developers, OS providers, and notification senders and at the outer edge, the institutional forces of employers, researchers, hardware makers, and policy bodies. Together, these layers show that attention is not just a personal resource, it is a systemic one.

Change and Impact

This comparison highlights the shift from interruption-blind digital systems to attention-aware design. Current notification architectures are built around the sender’s intent, not the receiver’s cognitive state. A message is delivered the moment it is sent, regardless of whether the recipient is mid-task, in a flow state, or already cognitively overloaded. The result is a system optimized for immediacy at the cost of sustained focus.

Inclusion & Accessibility

Access to an attention-aware system is not uniform. The ability to benefit from interruption management depends on a range of cognitive, physical, and social factors that are unevenly distributed across users. Some barriers are internal: a person with ADHD may experience interruption recovery very differently from a neurotypical user: for them, even a brief disruption can cascade into a much longer loss of focus. Users in high-stress or high-anxiety states are more vulnerable to the compounding effects of notification overload. Inclusion here is about cognitive and emotional accessibility, not just physical or perceptual access.

New semester, new researches and new problems…Moving from theoretical research into lo-fi prototyping forced me to confront something I had been putting off: my topic is hard to make it physical. Research on attention, interruption and cognitive recovery does not translate naturally into a clickable mockup or a paper sketch. The phenomenon I am studying is internal and time-dependent. You cannot see it on a screen.

The approach I took was a simple behavioural task. A participant reads a short text and corrects errors in it while thinking aloud. I interrupt them once with a spoken question, then observe how they return to the task. To capture focus levels over time, I built a small browser tool; a slider the participant adjusts continuously, which logs their self-reported focus every thirty seconds and marks the exact moment of the interruption. At the end of the session it exports a CSV, so it is easier to track the data. It is low-tech, but it produces real data grounded directly in the literature.

The first test format was a speed dating exercise: five-minute exchanges rotating between participants. The idea is to get fast, varied feedback, and it works well for many types of prototypes. For mine, it was harder to make useful. Five minutes is not long enough for a person to actually get into the task before the interruption happens. Most of the time was spent explaining the setup rather than experiencing it, so the feedback reflected how clearly I could describe the concept more than how the prototype itself worked.

The most consistent piece of feedback was about the setup: having the task on one device and the focus slider on another felt fragmented. Several people pointed out that if I am studying attention, adding friction between two separate tools is counterproductive. It is a fair observation and one I plan to address, consolidating everything into a single interface makes more sense both practically and conceptually.

The session was useful in a different way than I expected. It did not tell me much about interruption and recovery, but it did clarify the limits of the current prototype as a communication tool. My research sits closer to the theoretical end of the spectrum for now, and that makes it genuinely difficult to prototype in a format that works for quick explanation to someone unfamiliar with the underlying psychology.

The plan is to run the full session individually, without the time pressure, in a setting where the participant can actually reach a state of focus before the interruption lands. I also want to think about whether a behavioural experiment alone is the right form, or whether the research needs a more designed artefact alongside it, something that makes the concept visible rather than just measurable.