MouthPad^ Benchmark (Augmental, 2026)
Measured by Augmental (Tomás Baptista) · MouthPad^ Benchmark Run (2026)
Inputs
The measured or assumed values behind the calculations, each with its source.
- TP = 3.53 bits/s
- On-screen MouthPad^ Benchmark value labeled BPS (Fitts), the benchmark's own Fitts throughput for the run. The reference calculation re-derives this from the grid geometry and the headline BPS and reproduces it (~3.5 bits/s), so Augmental's displayed Fitts value serves as a cross-check rather than a face-value input.
- B = 8.51 bits/s
- Headline MouthPad^ Benchmark score in a 60-second run by Tomás Baptista using head-tracking and tongue clicks (YouTube title/description and on-screen result).
- grid = 30x30
- On-screen benchmark grid size; the final scoreboard also shows 32 hits and 100% success.
Strictest ITR
Each scoring method is an upper bound on the channel, so the headline is the strictest (smallest) one for this entry. Use the score selector on the home page to view any single method across entries.
-
Grid geometry -> movement difficulty
30x30 grid: W = S/30; random targets -> mean amplitude A ~= 0.52*S, so A/W ~= 15.6. ID = log2(A/W + 1) = log2(16.6) ~= 4.05 bits/movement.
Same grid correction used for Neuralink Webgrid and the BrainGate2/Card grids: a cued target is a pointing movement, so the information is the Fitts index of difficulty (~4.05 bits), not the log2(900) the full-grid score credits.
-
Augmental's displayed Fitts value, cross-checked against the headline BPS
Benchmark shows 'BPS (Fitts)' = 3.53 directly. Cross-check by re-crediting the headline 8.51 BPS: 8.51 × 4.05 / log2(899) = 8.51 × 4.05 / 9.81 ~= 3.5 bits/s, matching.
Unlike Neuralink's Webgrid, Augmental displays its own Fitts value, so 3.53 is taken from the benchmark and the re-credit of the headline score confirms it rather than standing in for it. ITR = 3.53 bits/s.
What counts as a bit depends on the action space. The number of distinguishable actions and how likely each one is are design choices of the task, not the sensing hardware. The same modality can present a fixed set of targets, a set pruned per step by a grammar or language model, or a continuous control space. Each of these changes how many actions are live and how the probability mass is spread, and therefore the information per selection. Read the action space below before comparing headline numbers across entries.
Action space
What the user can produce at each step, and how those options are distributed.
- Structure
- Continuous control space
- Size
- Continuous
- Prior
- Uniform: all actions assumed equally likely
- Notes
- A hands-free cursor-control channel using head tracking for pointer motion and tongue/sip gestures for commands. The benchmark presents a cued 30x30 target grid. Augmental reports both the full-grid BPS headline and a separate Fitts BPS value; the Fitts value is used as the reference because it measures the continuous pointing channel in the same terms as mouse, trackball, stylus, tongue-drive and cursor-BCI entries.
Comparability The strictest bound here is Fitts throughput: the index of difficulty, log₂(A/W + 1), per movement. Directly comparable to the other continuous-pointing entries (mouse, trackball, stylus, gaze and the cursor BCIs). Set against the text entries (keyboards, spellers, speech) it crosses methods: both report bits/s, but one measures movement difficulty and the other text information, so compare within the family first.
Other score types
Bounds the atlas keeps out of the default strictest headline: as-reported figures, alternate task conditions, or raw-channel ceilings that shouldn't win the headline by default. Each still carries a score type, so the home-page selector ranks this entry on it when you choose that type. Read its derivation before comparing across entries.
-
Achieved-bitrate metric
30x30 grid -> N = 900 cells; the Webgrid-style score credits log2(N - 1) = log2(899) ~= 9.81 bits per net-correct cued selection.
-
Augmental's reported full-grid score (taken as reported)
BPS = 8.51 bits/s on the 30x30 grid (the description also notes a 9.98 BPS personal best, not the run shown).
Reported on-screen from a 60 s run at 100% success; the benchmark gives only the BPS scores, not a separate selections/s, so this is taken as reported. Dividing by the 9.81-bit credit implies ~0.87 selections/s, a consequence of the score rather than an independent measurement.
Source
- Authors
- Augmental (Tomás Baptista)
- Publication
- MouthPad^ Benchmark Run, 2026
- Reference
- Augmental MouthPad^ product page
- Reference
- MIT News profile of MouthPad mechanics