A few weeks ago, Professor Síle O’Modhrain’s new book landed on my desk: A Dialogue of the Senses: Designing Interaction with the Body in Mind, out this spring from MIT Press. I’d known it was coming — she’d mentioned it once, almost in passing, the way you mention something you’ve been working on for years. What I didn’t expect was how directly it would end up shaping the next phase of Run Like a Fish.
I read it looking for background. I came away with a testing plan.
The idea that stuck with me
The book’s central argument is that we don’t design experiences for people — we design the conditions under which people, by moving their bodies, bring an experience into being themselves. That sounds abstract until you apply it to a wearable that’s supposed to guide a blind runner down a path. It means the wearable isn’t “delivering navigation information” the way a screen delivers a map. It’s part of a loop: the runner moves, their body picks up a cue, they adjust, and the meaning of that cue only exists in the moving and adjusting. You can’t design the experience directly. You can only design good conditions for it.
Three chapters in particular reframed how I was thinking about my own research: the ones on time, movement, and space. The short version of what I took from them:
- Haptic design isn’t just about what information you send someone. It’s about when you send it — timing changes meaning as much as content does.
- Movement itself changes perception. A signal that’s crisp and obvious when someone is standing still can become unreadable the moment they start running.
- People don’t understand a new spatial language instantly. They build it gradually, cue by cue, over repeated exposure — which means a testing plan has to leave room for learning, not just measure first-try accuracy.
I’d been circling all three of these ideas already, in a scattered way. The book gave me the vocabulary to make them the actual structure of my next round of studies.
Rebuilding my testing plan around Time, Movement, and Space
Here’s where it landed: instead of one big test of “does the haptic wearable work,” I’m running a set of Wizard-of-Oz studies organized around exactly those three dimensions.
Quick explanation for anyone new here — Wizard-of-Oz testing is a method where, instead of building an autonomous device first, a person (the “wizard”) manually triggers the feedback while the runner experiences it as if the device were deciding on its own. I get to test how a runner responds to guidance — timing, intensity, how it feels in motion — without first solving sensing, computer vision, or autonomy. The wearable can be fake. The interaction has to be real.
Under that method, the three chapters became three concrete lines of testing:
Movement. How should feedback change once someone is actually running, versus standing still? I’m comparing continuous guidance (a constant stream of correction) against discrete guidance (only cueing when something needs to change), and static feedback against feedback that adapts with pace.
Time. When should a cue arrive, and how often? I’m testing early cues against late ones, high-frequency against low, and constant reminders against guidance that only shows up for major decisions — measuring reaction time, missed turns, and how much mental effort each option costs.
Space. How should information actually sit on the body? This is body placement (wrist, forearm, waist, ankle), how rich a signal vocabulary someone can actually hold onto, and whether a vibration’s location naturally reads as a direction without having to be taught.
Layered into all three is something I’ve been excited about for a while: comparing a handful of biomimetic guidance strategies head-to-head inside these same studies — a lateral-line-inspired continuous pull toward center, a bat-echolocation-inspired pulse that speeds up as a turn approaches, and an optic-flow-inspired drift correction. Rather than committing to one of these up front, the Wizard-of-Oz platform lets me test which one actually feels intuitive before any of it becomes hardware.
Why this matters more than it might sound like it does
It would have been easy to read the book, nod along, and keep my original plan. What changed my mind was realizing my original plan measured whether people could detect signals — full stop — without asking whether they could detect them while moving, or fast enough, or after enough repetition to actually trust them. Time, movement, and space aren’t just nice organizing labels. They’re the three ways a “successful” signal in a quiet lab can quietly fail once someone’s actually running with it.
I’ll be walking Professor O’Modhrain through this restructured plan soon, and I’m genuinely looking forward to hearing where she thinks I’ve got it right — and where I haven’t.
More soon.
This post is part of an ongoing research series at runlikeafish.blog exploring biomimetic approaches to assistive technology for visually impaired runners. If you’re a visually impaired runner, a guide, or a researcher working in this space, I’d love to hear from you. You can reach me at brooke@runlikeafish.blog