ITR Atlas
← All interfaces

Neurotrophic Electrode (Kennedy & Bakay, 1998)

Measured by Kennedy & Bakay · NeuroReport 9(8) (1998)

Intracortical Cursor invasive 1998

Inputs

The measured or assumed values behind the calculations, each with its source.

rate = 3 char/min
Patient JR (Johnny Ray, locked-in) drove a 2D cursor over a virtual keyboard and selected characters via imagined hand movements, producing ~3 correct characters/min (Brumberg, Nieto-Castanon, Kennedy & Guenther 2010; see references). This is the documented real-world typing rate.
H = 1.0 bits/char
English-text entropy (Shannon); the same ~1 bit/char standard used for QWERTY, eye-typing, Synchron and the BrainGate2 typing entry.
N = 28
≈ alphabet size of the virtual keyboard JR pointed at, for the raw-key Wolpaw ceiling (uniform prior). Approximate; the source documents the rate, not an exact key count.
T_key = 20 s/key
Key-selection interval for the Wolpaw ceiling: 60 / 3 char/min = 20 s. No per-selection accuracy is reported, so the bound is taken at perfect copy (P=1) as a strict ceiling.
signals = 3
Three neural signals mapped to cursor x, cursor y, and a binary select (Kennedy & Bakay 1998; Kennedy et al. 2000).

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.

Strictest Shannon (text) Recomputed
Character-entropy throughput (realized text entry)
Net of English redundancy
0.05 bits/s

  1. Characters per minute

    ≈ 3 correct char/min (patient JR: 2D cursor over a virtual keyboard, imagined hand movements)

    The documented real-world typing rate for this first chronic intracortical communication BCI. The cursor supplies the pointing; characters are the realized output.

  2. Bits per character

    H(English) ≈ 1.0 bit/char (Shannon)

  3. Information transfer rate

    3 char/min × 1.0 bit/char ÷ 60 s/min = 0.05 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
Context-conditioned: likelihoods depend on prior actions
Notes
The first chronic intracortical communication BCI (patient Johnny Ray, locked-in). A low-DOF continuous cursor plus a select signal, used to point at characters on a virtual keyboard, i.e. pointing-to-type, like the BrainGate2 typing and Synchron entries. The realized output is English text, so the reference uses the character-entropy method (~1 bit/char) on the documented 3 char/min rate rather than a raw cursor bitrate. Independent reviews quote ~0.03-0.05 bits/s for this system, which agrees with the 0.05 bits/s derived here.

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 bounds considered for the headline

Also valid upper bounds for this entry and eligible to be the headline. They just came out looser than the strictest above. Pick any of these in the home-page score selector.

Wolpaw Recomputed
Wolpaw bitrate over the raw key set
Uniform-prior, perfect-copy ceiling on the key channel
0.24 bits/s

  1. Bits per selection (Wolpaw formula)

    B = log2(N) + P*log2(P) + (1-P)*log2((1-P)/(N-1))
  = log2(28) + 1*log2(1) + 0*log2(0/27)
  = 4.807 bits / selection

    Term 1 is the information if every choice were correct; terms 2-3 subtract the bits lost to the error rate, assumed spread evenly over the other N-1 targets.

  2. Selections per second

    T = 20 s/selection  ->  1 / 20 = 0.05 selections/s

  3. Information transfer rate

    ITR = B * selections/s = 4.807 * 0.05 = 0.24 bits/s

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.

Nuyujukian Recomputed
Nuyujukian achieved bitrate over the raw key set
Achieved-bitrate view of the key channel, shown for comparison
0.24 bits/s

  1. Achieved-bitrate credit per net-correct key

    N = 28 keys → log2(N − 1) = log2(27) = 4.75 bits per net-correct selection (field-standard achieved bitrate, e.g. Webgrid; Nuyujukian 2015, which introduced the metric, used log2(N)).

  2. Net-correct key rate

    No per-selection accuracy is reported, so this is a perfect-copy ceiling: net-correct = the full 3 char/min = 0.05 correct/s (one key per 20 s).

    Same N and rate as the entry's perfect-copy Wolpaw ceiling. With no error data the achieved and Wolpaw ceilings differ only by log2(N − 1) vs log2(N), so both land at ~0.24 bits/s, above the 0.05 bits/s Shannon headline.

  3. Achieved bitrate

    4.75 bits × 0.05 correct/s = 0.24 bits/s.

Source

Authors
Kennedy & Bakay
Publication
NeuroReport 9(8), 1998
Paper
10.1097/00001756-199806010-00007
Reference
Brumberg, Nieto-Castanon, Kennedy & Guenther 2010 (Speech Communication): reports JR produced ~3 characters/min over a virtual keyboard
Reference
Kennedy et al. 2000 (IEEE Trans. Rehabil. Eng.): fuller 3-signal cursor-and-select control report