Booth Id:
ENBM074
Category:
Biomedical Engineering
Year:
2026
Finalist Names:
Kharade, Ameya (School: Nashua High School South)
Abstract:
Your mind works perfectly, but your body will not let you speak. Millions with Amyotrophic Lateral Sclerosis, stroke, or paralysis live trapped inside working minds. Today's brain-computer interfaces (BCIs) manage 5 words-per-minute (WPM): forty seconds to say "I love you" while the moment passes. Implants reach real speeds but require brain surgery and cost over $100,000.
I discovered a new paradigm for brain-machine communication: resolving intent directly, instead of spelling messages out letter by letter. I validated this using $1,800 of consumer electroencephalography hardware and compared it against conventional decoding across 111 randomized, counterbalanced trials on identical hardware, with all effects independently replicated on a larger naive cohort.
My system surpassed the field's 30-year ceiling of 15 WPM, achieving 65 WPM versus 3 WPM on a conventional speller — a 20-fold improvement — with 23-fold lower
time-per-message (10.7s vs 242.2s, q<0.001) and 43-fold intent-amplification (Cohen's d=2.9, q<0.001), a new metric quantifying meaning conveyed per brain signal. A double dissociation then confirmed fundamentally different mechanisms: traditional effort scales with message length (ß=0.72, q<0.001) but is blind to intent ambiguity (ß=-0.02, q=NS), while my system is blind to length (ß=0.28, q=NS) and scales with ambiguity instead (ß=0.45, q<0.01). Speed now depends on how ambiguous a thought is, not how long the sentence is.
The Six Laws of Intent Resolution I derived from information theory formalize this for any bandwidth-limited interface — broadly reshaping biomedical engineering, and making every existing BCI, prosthetic, and beyond, dramatically more efficient. A new foundation for how humans and machines communicate, and a voice for the voiceless.
Awards Won:
Second Award of $2,400