NeuroNoise — The Future of Attention with BCI, Smart Glasses, and AI

Original Post: NeuroNoise — The Future of Attention with BCI, Smart Glasses, and AI

Improved Attention for Improved Quality of Life

Why Attention Matters

Attention is vital for a variety of cognitive processes. The first step of forming memories is encoding, and inattention limits encoding. Similarly, if an individual is focused on the wrong things, they may fail to encode and store the memories they wish to retain. Memory is just one area of impact — attentional deficits also affect our ability to learn, socialize, and attain our goals. Psychological research has revealed that focus in states of flow and meditation are prime indicators of individual long-term happiness (Killingsworth and Gilbert 2010).

Given the importance of attention, it’s concerning that impaired attention is a symptom of many common diseases, like dementia, ADHD, OCD, and depression. Further, modern life presents individuals with a stream of distractions (social media, urban environments, etc.) that impair attention in a variety of forms — low attention span, poor selective attention, high distractibility, etc. Thus, solutions which help mitigate attentional impairments and train our natural ability to focus could have a drastic and positive impact on people’s lives. My own interests lie in using technology to enhance memory, communication, and mental health. I believe that a system to help people attend can drastically improve our ability to remember, communicate, and flourish.

Existing Attentional Enhancement

A number of techniques and solutions exist to improve our attention. Mindfulness meditation is a proven technique and benefits from a low barrier to entry. However, mindfulness meditation can be a difficult practice to maintain and the lack of feedback can make it difficult to perform correctly. Another solutions lies in pharmaceuticals which enhance classical measures of attention but often lead to dependence and negative side effects. Noise cancelling headphones or white noise speakers are excellent methods to avoid distraction, but they don’t allow individuals to selectively avoid cancelling out the information they want to attend to. Finally, neurofeedback methods for attention, specifically for ADHD, have been shown to be effective (Arns et. al. 2013). However, these interventions requires dozens of sessions to see effects, and don’t help with in-the-moment attentional needs. An attentional modulation and training system is needed that helps users attend to the right things and maintain focus, in a way that is easy to adopt in daily life and that provides real-time and closed-loop feedback, while avoiding dependency, side effects, and high barrier to entry.

Closed-loop, Real-time Attentional Modulation

I propose a closed-loop, real-time attentional modulation system which senses and acts on individuals’ neurological response to incoming sensory information and internal processing of semantic data. The system will sense what a user is attending to and pair this with a “goal focal point” — the media, person, information, or idea that the user wants to attend to. The system then guides the user to attend to their “goal focal point” through a variety of interventions. The system would train users to enter a state of focus, ignore distractions, and maintain attention. Increased attention on the “goal focal point” can have immediate impact on the ability of users, especially those with memory impairments, to focus on, encode, and store memories, alongside benefits to learning, socialising, communication, mental health, and more.

“Metasomatic Mind Maps “— Attentional Meditation Training

An example of this type of system could train users’ to perform body scans, a metasomatic meditation technique shown to be beneficial in training attention (Garcia et. al. 2018, Morone et. al. 2008). The system, called “Metasomatic Mind Maps” (Jain et. al. 2020) could use EEG of the parietal lobe to map user’s interoceptive attention to body location when performing a body scan (Mondini et. al. 2021, Schaefer et. al. 2002). The system could identify the somatic attention of the user and provide live feedback in terms of haptic sensations on the body location of focus, or a live avatar with a somatic attention heatmap overlaid on its body, in order to help users learn to attend through body scans.

“Listen Here” — Disengagement Detector

Another system could help user’s monitor and improve their attention during conversations or other spoken content. A system, called “Listen Here!”, could constantly sense how much attention an individual is paying to a conversation, and provide cues when attention fades. This could be achieved through sensing of auditory evoked potentials and speech (Myers et. al 2019). The cues could be private (e.g. a subtle haptic pulse, or indicator light flash in the corner of users’ eyes’) or public (e.g. a shared cue audio cue). Similarly, if someone is watching a training video or live lecture, the system could understand when the user’s visual attention decreases and deliver a cue. This visual attention tracking could be achieved by correlating user’s visually evoked potentials with their visual environment.

“One Thing at at Time: Selective Attention through Diminished Reality”

Often, an individual wants to focus on some signal while distractions happen in the background, which can be difficult to ignore. For example, when two friends in a coffee shop are trying to discuss something important, but there are loud tables nearby. A selective noise cancelling system, called “One Thing at at Time: Selective Attention through Diminished Reality”, could accept a “goal focal point” and selectively block audio and visual information from anything but the “goal focal point”. This could be achieved with a pair of earbuds for audio and/or AR glasses for video. User’s could specify this “goal focal point” by pointing their finger or taking a quick picture of whatever they wish to attend to.

“NeuroNoise” — Brain Stimulation Sensory Attentional Modulation

To combat dependence on noise cancelling aids, a brain stimulation system, called “NeuroNoise”, could actively diminish the user’s neurological response to distracting sensory signals, shifting their attention away from those signals. This could be used to train users what it feels like to ignore incoming sensory data. Stimulation can be removed smoothly as the user independently takes on the task of selective attention. Similarly, neurological responses to sensory information could be enhanced, helping users learn what it feels like to focus on incoming information, and gradually rolling back the intervention so users learn to attend without the use of external aids. This could be achieved using environmental sensors and tACS/RNS to increase or decrease sensory evoked potentials (Erkens at. Al 2021, Wang et. al. 2020). The focus on training and state switching over sustained states allows for a system which helps users achieve their full potential without external aids, and avoids problems associated with long term use of tES systems.

“Stream of Conscious Consciousness” — Semantic Brain Sensing

Brain sensing smart glasses

A common issue for those with ADHD, OCD, and other attentional impairments is selective attention, or the ability to identify and maintain a focal point. For example, an individual might wish to sit down and brainstorm ideas for their next essay about food sustainability, but find their mind constantly wanders to the football game on tonight. A system, called “Stream of Conscious Consciousness”, could accept a current “goal focal point” from a user, and notify the user whenever their internal thinking strays too far from that focal point, helping users train their selective attention while still allowing flow between related ideas. This live tracking of the general semantic category of a user’s thoughts could be achieved with semantic decoding using mobile EEG (Simanova et. al. 2010, Kosmyna et. al. 2023, Correia et. al. 2015) in an all-day wearable glasses or baseball cap form factor.

“Latent Lane” — Semantic Brain Stimulation

A next step in this direction would take this semantic decoding and pair it with semantic brain stimulation. This system, called “Latent Lane”, could actively inhibit memories/thoughts about concepts that the user specifies as distracting, or activate the right memories and boost users’ thinking in a particular semantic direction, helping them focus on the right concept (Jacobs et. al. 2011). Instead of merely being notified when their thoughts have strayed from food sustainability to football, the system can actively inhibit thoughts of football to help the user stay on topic. While still difficult to achieve today, recent advances in BCI semantic decoding, the reciprocity principle (Fernández-Corazza et. al. 2016), and ML encoders trained on BCI datasets could be combined in a system to influence user’s thinking in a specific semantic direction that the user actively defines and controls.

“Think With Me” — Conversational AI Remains on Topic

Many people find that thinking aloud can help them think things through and stay focused on a topic. A conversational AI avatar, called “Think With Me”, can focus on providing simple responses, with the goal of helping the user stay focused on a particular topic. The system could continually monitor the semantic content of what’s being discussed and, when the conversation strays, generate a comment or question with the goal of bridging the conversation back to the “goal focal point”. This system avoids any need to sense or stimulate the user neurologically by simply helping the user stay on track in their thought process.

Meta-attention — What Should We Attend To?

While the systems discussed thus far have assumed a user-defined “goal focal point”, these systems could also recommend to the user what they should attend to. This can be especially useful for those with OCD, ADHD, and depression, who struggle with misplaced attention. The system could identify the information that is most useful to an individual given their current context and goals, recommending a “goal focal point” — it’s then up to the user to choose a recommendation, or define their own. In this case, the human in the loop maintains full autonomy, and the AI acts to augment human attention, not only helping us attend, but attend to the right things.

Conclusion

In conclusion, I envision, and want to create, an attentional modulation system that helps teach people to enter into a status of focus, maintain focus, and to practice selective attention towards whatever matters to them. This system has the potential to improve quality of life for people with attentional and memory deficits, as well as train everyone to improve their ability to pay attention to what matters.

Acknowledgements

Thank you to Tomás Vega for ideas, edits, and encouragement throughout writing this piece.