Neurotech

Source: “Reconstruction encompasses 450 X 350 X 52 um volume of tissue.” Derived geometric volume: 450 × 350 × 52 / 10⁹ = 0.00819 mm³. Publication-year checkpoint; first release date not established. Dense reconstruction of every neurite is not asserted.

Source: “The hemibrain sample is roughly 250 × 250 × 250 μm.” Geometric product 0.015625 mm³, displayed approximately 0.0156; not acquired voxel-mask volume. Publication year, not exact release date.

Source: “roughly one cubic millimeter of brain tissue.” Announcement dated June 1, 2021. Automated segmentation and annotated synapses; only selected cells proofread. Later 1.05-mm³ compression-corrected paper estimate is not new tissue.

Source: “first public data released on the cubic millimeter dataset in July 2021.” Approximate nominal sample scale, not an exact valid segmentation mask. July precision; 2025 paper is later analysis of this specimen.

Source: “The whole brain contains 0.0175 mm³ of neuropil volume.” Publication-year checkpoint for this reported value; the reconstruction reuses FAFB imagery and is not new tissue acquired in 2024.

Figure 1: “0.95 × 10⁶ µm³ before” expansion, equivalent to 0.00095 mm³. Native volume, not expanded hydrogel; automated dense reconstruction with synaptic molecular predictions and selected traced validation regions. Publication-year checkpoint, first final-dataset release not established.

Preprint: “0.082 mm3 total volume” (160 teravoxels at 8-nm isotropic resolution). Release date independently verified at https://male-cns.janelia.org/ : “2025-10-03 - MaleCNS v0.9 released”. Later v1.0 and 2026 publication reuse this specimen; not all synapses have proofread partners.

Source: “These groups have implanted a total of 67 participants”; data collection ended December 2023 and later implants are excluded. Includes ECoG and endovascular systems, not only intracortical. Only 31 were active at cutoff. Supplement Table 2 identifies participants but does not supply implantation-year columns; no annual points invented.

April 12, 2024 disclosure: “In January, we conducted the first human implantation.” January is the implantation month, not an exact surgery day.

August 21, 2024 disclosure: “Last month, Alex ... the second participant ... received his Neuralink implant.” July implantation month; cumulative two recipients.

Source: “Today, there are three people with Telepathy: Noland, Alex, and Brad.” Disclosure date, not all three surgery dates.

Source: “Five individuals with severe paralysis are now using Neuralink to control digital and physical devices with their thoughts.” Post-implant use supports receipt, not mere enrollment.

Company headline says “21 Neuralnauts enrolled”; implantation interpretation also requires body: “higher signal quality across 18 of the subsequent 20 participants” after the first recipient. Company-reported cumulative checkpoint, not independently audited current active users. Page release date Jan 28; last modification Feb 2 is not the checkpoint.

Issuer: “Stentrode BCIs have been placed in 10 patients with paralysis to date, across clinical trials in the U.S. and Australia.” Includes four Australian and six US recipients; do not add those subsets or add this to review baseline.

Source: “Seven study participants have now received ARC-BCI Therapy.” Two additional people bring total to seven. Some historical CEA/WIMAGINE participants overlap the review; exact crosswalk unresolved.

June 17, 2026 issuer release: “The first participant ... was enrolled and implanted ... and will be followed and evaluated over the next six years.” Exact disclosure date, not surgery date. At least one long-term participant; earlier temporary 2025 epilepsy implantation is excluded.

Table 1 reports v1.0.0, 2017-04-17, Total Duration (Hours): 170. Full corpus size, not seizure-event duration or channel-hours. Registration-controlled data; evaluation partition withheld in accompanying poster. Versions overlap and must not be summed; unchanged totals can accompany annotation revisions.

Table 1 reports v1.1.0, 2017-08-04, Total Duration (Hours): 425. Full corpus size, not seizure-event duration or channel-hours. Registration-controlled data; evaluation partition withheld in accompanying poster. Versions overlap and must not be summed; unchanged totals can accompany annotation revisions.

Table 1 reports v1.2.0, 2018-04-15, Total Duration (Hours): 504. Full corpus size, not seizure-event duration or channel-hours. Registration-controlled data; evaluation partition withheld in accompanying poster. Versions overlap and must not be summed; unchanged totals can accompany annotation revisions.

Table 1 reports v1.3.0, 2018-08-16, Total Duration (Hours): 651. Full corpus size, not seizure-event duration or channel-hours. Registration-controlled data; evaluation partition withheld in accompanying poster. Versions overlap and must not be summed; unchanged totals can accompany annotation revisions.

Table 1 reports v1.4.0, 2018-11-14, Total Duration (Hours): 651. Full corpus size, not seizure-event duration or channel-hours. Registration-controlled data; evaluation partition withheld in accompanying poster. Versions overlap and must not be summed; unchanged totals can accompany annotation revisions.

Table 1 reports v1.5.0, 2019-07-22, Total Duration (Hours): 1,074. Full corpus size, not seizure-event duration or channel-hours. Registration-controlled data; evaluation partition withheld in accompanying poster. Versions overlap and must not be summed; unchanged totals can accompany annotation revisions.

Table 1 reports v1.5.1, 2020-04-23, Total Duration (Hours): 1,074. Full corpus size, not seizure-event duration or channel-hours. Registration-controlled data; evaluation partition withheld in accompanying poster. Versions overlap and must not be summed; unchanged totals can accompany annotation revisions.

Table 1 reports v1.5.2, 2020-05-09, Total Duration (Hours): 1,074. Full corpus size, not seizure-event duration or channel-hours. Registration-controlled data; evaluation partition withheld in accompanying poster. Versions overlap and must not be summed; unchanged totals can accompany annotation revisions.

Release metadata: “Neural recordings are available at 500 Hz from at least 64 electrodes per participant, for a total of 1280 hours.” datePublished January 27, 2022. Reported duration, not electrode-multiplied time; gaps/preprocessing mean nominal and QC-valid hours differ.

Abstract: “over 100 hours of recordings” from seven nonhuman primates and over 158 sessions. Strict lower-bound marker, not exactly 100. Existing datasets reused for training; no new-data or wholly open-corpus claim.

Abstract: “1020 hours of simultaneously recorded scalp electroencephalography (EEG), facial electromyography (EMG), and speech audio”; authors say released through OpenNeuro under CC0. Descriptor submission date, not verified archive upload date. Count once, not three modalities; full manifest not enumerated.