DANDI:001754

Reproducing Knierim et al. (2000) from dandiset 001754 10

Notebook reproduce_knierim_2000.ipynb

• ✕Not reproducedMain text, paragraph beginning "The mean spatial-information content for all 7", sentence 2; Figure 3a,b and legend

  For rat 3 on flight day 4, most of the 7 cells kept the same firing fields in a third session that followed the second immediately; the fourth cell lost its field, and a few other cells gained a field in session 3.

  Another recording session followed immediately, and most cells maintained the same firing fields in both sessions (Fig. 3b), demonstrating that the spatial tuning was stable across different exposures to the track.

  Main text, paragraph beginning "The mean spatial-information content for all 7", sentence 2; Figure 3a,b and legend

  Under M-b, the one reading that can be run, 6 of 16 active session 2 units keep their field by reading (a) (same-cell correlations -0.024 to 0.366, 95th percentile of 240 different-cell pairs 0.094) and 1 of 16 by reading (b), neither more than half (8), so the rule gives not reproduced. M-a is not run, as the criterion says when its pairing condition fails. Each distinct deposited map with a unit on its tetrode, 17 of 19 in session 2 (all 11 active maps among them) and 23 of 23 in session 3, was tied to the unit of its tetrode whose spike-count map it best matches. In session 2 no unit is tied to two maps (two maps are each the best map of two units, the second unit matching at r 0.009 and 0.059); in session 3 four are (TT10 cluster 6, TT4 cluster 3, TT5 cluster 2, TT8 cluster 5), three of them units of active session 2 maps. Session 2 has 16 active units and 11 active deposited maps (mean rates 0.061 to 2.500 Hz), not 7, and nothing in the files marks which 7 the paper showed. Reading (b) under M-b gives 2, 1, 0 and 0 cells at field thresholds of 0.1, 0.2, 0.3 and 0.5. Four units below 0.05 Hz in session 2 are active in session 3, which fits the legend's cells that gained a field; which cell lost its field is not judged. What changed. Before round 1 of verification the check tied each unit to its best map and silently kept the better match when two units chose one map, so it did not test the criterion's condition; it passed, M-a was judged, and its reading (a) kept the claim blocked. The answers of 2026-10-20 had made "most" more than half of the active cells, not a count of 4. The outcome differs from 001754-knierim-2000-02 (reproduced), which used a median correlation of smoothed maps that the paper does not give; the claims file now gives a per-cell rule for 7 cells.

  How the test differed from the paper’s method

  • Session 2 and session 3 are read as the file's ES1 and ES2.
  • Reading M-a is not run, as the pairing of deposited maps to units by best match, which the criterion allows only when one to one, is not one to one; the dandiset gives no link. The claim rests on unsmoothed recomputed maps with few spikes per 4-pixel bin.
  • Takes the same tetrode and cluster id in both epochs to be the same cell.
• ✕Not reproducedMain text, paragraph beginning "It is interesting to note", sentence 2; also the standfirst

  Hippocampal cells formed distinct, reliable representations of position on the three orthogonal surfaces of the Escher staircase in microgravity.

  Thus hippocampal cells can form unique, reliable representations of position on three orthogonal surfaces in microgravity

  Main text, paragraph beginning "It is interesting to note", sentence 2; also the standfirst

  Part (i) fails, as fig3-stability-across-sessions is not reproduced, and the criterion gives not reproduced when either dependency is not reproduced, whatever part (ii) gives. flight-spatial-code is blocked as methods_underspecified, which does not change this. Part (ii) could not be tested in any case, as it needs the three surfaces drawn in the position frame and the dandiset holds one two-dimensional pixel position with no camera view, track layout or pixel scale; the NASA record says two synchronised cameras recorded position, and the conversion code at b538d9e describes the track only as three-dimensional with 90 degree turns. Before round 1 of verification both dependencies were blocked, and so was this claim; it moved with fig3-stability-across-sessions. In 001754-knierim-2000-02 it was blocked as metadata_insufficient.

  How the test differed from the paper’s method

  • Inherits the deviations of fig3-stability-across-sessions (judged on M-b alone).
• ◐Partially reproducedMain text, paragraph beginning "Spatial firing patterns of 16 active place cells", sentences 1 and 3; Figure 2a,b

  For rat 2, the spatial information per spike of the 16 active cells recorded on the Escher staircase on flight day 9 (1.13 +/- 0.09 bits) did not differ significantly from that of its 19 active preflight cells (1.12 +/- 0.08 bits), by Mann-Whitney test.

  An index of spatial tuning specificity ... did not significantly differ between cells recorded preflight and those recorded on FD9 (for active cells, defined as having a mean firing rate > 0.05 Hz on the track; mean information per spike ± s.e. preflight, 1.12 ± 0.08 bits, n = 19; FD9, 1.13 ± 0.09 bits, n = 16; not significant by Mann-Whitney)

  Main text, paragraph beginning "Spatial firing patterns of 16 active place cells", sentences 1 and 3; Figure 2a,b

  Under both readings with preflight values the Mann-Whitney test finds no difference. Rat 2's 17 active FD9 ES maps have 1.377 +/- 0.120 bits per spike, against 1.320 +/- 0.112 for its 21 preflight ES maps (p = 0.88, difference 0.057) and 1.284 +/- 0.116 for its 25 preflight MC maps (p = 0.66, difference 0.092). No reading puts both means within tolerance (1.377 above 1.22; 1.320 and 1.284 above 1.20), so the rule gives partially reproduced. The FD9 count (17) is within 16 +/- 1.6 and the preflight counts (21, 25) are not within 19 +/- 1.9. Reading M-b gives 2.965 +/- 0.224 over 25 active FD9 units but cannot be computed preflight; tested against the M-a preflight values it differs (p 6.2e-07 and 2.5e-07), a variant, not judged. Leaving out the maps with cell_number -1 (a variant) the test finds no difference either (p 0.87 and 0.56). The FD9 ES window is one in which the trace behaves as Rat 2's position (data check). In 001754-knierim-2000-02 this claim, then from the dandiset's description and judged on Rats 1 and 2 pooled by medians, was reproduced (later removed from that record); the outcome differs because the claim is now rat 2's, with the paper's means as expected values, which the deposited maps exceed. What the answers of 2026-10-20 changed. The criterion now has both sides of a comparison use one map reading and judges on M-a where M-b cannot be computed preflight, which is what this attempt had done; and under S-b each epoch of the preflight day is run on its own, so S-b gives the same two readings (ES and MC) as S-a, where the attempt had read S-b as MC alone. Neither changes a number or the outcome (partially reproduced before and after).

  How the test differed from the paper’s method

  • Reading M-a uses the deposited maps, each identical copy once, and counts a map as active by its own occupancy-weighted mean rate, as the pinned version cannot link maps to units.
  • Which preflight epoch was the rectangular track of Fig. 2b is not stated in the files.
• ◐Partially reproducedMain text, paragraph beginning "It is interesting to note", sentence 1 (given as unpublished observations)

  On the first experience of the Escher staircase on flight day 4, place cells of rats 1 and 2 showed abnormal patterns of spatial selectivity, which differed between the two rats.

  on the first experience on the Escher staircase on FD4, firing of place cells showed abnormal patterns of spatial selectivity that differed between rats 1 and 2

  Main text, paragraph beginning "It is interesting to note", sentence 1 (given as unpublished observations)

  Lower spatial information than preflight holds for rat 2 under both readings and for rat 1 under neither. Rat 2's 14 active FD4 ES maps (0.708 +/- 0.112 bits per spike) are lower than its preflight ES (1.320, p = 0.00063) and MC maps (1.284, p = 0.0015). Rat 1's 20 FD4 ES maps (1.409 +/- 0.192) are lower than its preflight ES maps without significance (1.804, p = 0.077) and higher than its MC maps (1.127, p = 0.36). Rat 2's FD4 maps are also lower than its FD9 maps (1.377, p = 0.00033); Rat 1's are not (1.588, p = 0.38). In Rat 2's FD4 ES window the trace could not be shown to be Rat 2's position (data check). The Rat1 and Rat2 files hold one trace there (4 of 707 shared valid positions more than 2 px apart), and Rat 2's FD4 units carry more information than every time shift on neither file's copy (0 of 17 and 1 of 16, against 12 of 28 and 16 of 28 on FD9), so whether the low FD4 information comes from firing or from tracking is not known; the criterion sets no condition on this. That file holds five identical copies of maps, counted once. Leaving out the maps with cell_number -1 (a variant) gives the same pattern (Rat 2 p 0.0002 and 0.0017; Rat 1 p 0.24 and 0.42). That the patterns differed between the rats is not judged. In 001754-knierim-2000-02 this claim, then from the dandiset's description and judged on Rats 1 to 3 pooled by a one-sided test, was reproduced (later removed from that record); it now requires each of rats 1 and 2 with a two-sided test, and holds for one. The answers of 2026-10-20 (one map reading on both sides, so M-b is not run as it has no preflight values; each preflight epoch a candidate) are what the attempt had done and change no number; partially reproduced before and after.

  How the test differed from the paper’s method

  • As for si-comparable-to-preflight.
• ✓ReproducedMain text, paragraph beginning "The mean spatial-information content for all 7", sentence 1; Figure 3a

  For rat 3, the mean spatial information of all 7 active place cells on flight day 4 (1.12 +/- 0.27 bits, session 2) was not different from its preflight cells (1.48 +/- 0.11 bits, n = 24).

  The mean spatial-information content for all 7 active place cells from rat 3 on the fourth day of flight (FD4) was 1.12 ± 0.27 bits (Fig. 3a), which was not different from the information content measured for this rat's place cells before the flight (1.48 ± 0.11 bits; n = 24).

  Main text, paragraph beginning "The mean spatial-information content for all 7", sentence 1; Figure 3a

  Under both readings the test finds no difference. Rat 3's 11 active session 2 (ES1) maps have 1.231 +/- 0.206 bits per spike, against 1.518 +/- 0.125 for its 38 preflight ES maps (p = 0.28, difference -0.287) and 1.337 +/- 0.121 for its 33 preflight MC maps (p = 0.77, difference -0.106). Under the ES reading both means are within tolerance (1.12 +/- 0.27 and 1.48 +/- 0.11), which the rule requires of one reading. The counts are not (11 against 7; 38 and 33 against 24), so the cells are not the paper's set; the criterion's rule does not use them. Session 3 (ES2, 19 maps, 1.384 +/- 0.177) does not differ from either preflight epoch (p = 0.28 and 0.86). Reading M-b gives 3.038 +/- 0.334 over 16 units, flight only, and is not judged. The answers of 2026-10-20 (one map reading on both sides; each preflight epoch a candidate under S-a and S-b) confirm the readings the attempt judged on and change no number; reproduced before and after. New claim.

  How the test differed from the paper’s method

  • As for si-comparable-to-preflight. Session 2 is read as the file's first ES epoch (ES1), as the paper says the first session was lost.
• ⊘BlockedMain text, paragraph beginning "During the Neurolab Space Shuttle mission", last sentence; supported by the paragraphs beginning "Spatial firing patterns of 16 active place cells" and "The mean spatial-information content for all 7" and Figures 2a, 2c and 3a; also the standfirst· Methods underspecified

  In flight, after an initial period, hippocampal place cells on the Escher staircase showed normal, spatially specific firing, with spatial information per spike not different from the same rats' preflight cells.

  Nonetheless, place cells eventually demonstrated normal, spatially specific firing properties.

  Main text, paragraph beginning "During the Neurolab Space Shuttle mission", last sentence; supported by the paragraphs beginning "Spatial firing patterns of 16 active place cells" and "The mean spatial-information content for all 7" and Figures 2a, 2c and 3a; also the standfirst

  The criterion judges this claim from its three dependencies. si-comparable-to-preflight is partially reproduced, si-rat3-fd4-vs-preflight reproduced and si-rat1-fd9-vs-preflight blocked as methods_underspecified. The others do not decide it, as the rule gives partially reproduced if the rat 1 comparison is reproduced or partially reproduced and not reproduced if it is not reproduced; the rat 1 preflight ES reading (p = 0.50) would give the first and the preflight MC reading (p = 0.0055, flight higher) the second. The outcome differs from 001754-knierim-2000-02 (not reproduced) because the claim is now judged as the paper judges it, by spatial information against preflight, and not by a shuffle test the paper does not use; that test is not run. The answers of 2026-10-20 leave the rat 1 comparison blocked, as its criterion now says when the two preflight epochs disagree, so this claim is blocked before and after.

  How the test differed from the paper’s method

  • Inherits the deviations of its three dependencies.
• ⊘BlockedMain text, paragraph beginning "Spatial firing patterns of 16 active place cells", sentences 4 and 5; Figure 2c· Methods underspecified

  For rat 1, the mean spatial information of its 21 active cells on flight day 9 (1.19 +/- 0.09 bits) did not differ significantly from its preflight cells (1.37 +/- 0.10 bits, n = 28).

  We also determined the spatial tuning of 21 active cells from rat 1 on FD9 (Fig. 2c). ... did not significantly differ from that rat's preflight data (1.37 ± 0.10; n = 28).

  Main text, paragraph beginning "Spatial firing patterns of 16 active place cells", sentences 4 and 5; Figure 2c

  The readings disagree on the test. Rat 1's 24 active FD9 ES maps (1.588 +/- 0.143 bits per spike) do not differ from its 25 preflight ES maps (1.804 +/- 0.149, p = 0.50, difference -0.216) and are higher than its 24 preflight MC maps (1.127 +/- 0.114, p = 0.0055, difference 0.461); the paper does not say on which track or in which epoch rat 1's preflight cells were recorded. No mean is within tolerance (1.19 +/- 0.09, 1.37 +/- 0.10) and no count within 10% (24 against 21; 25 and 24 against 28). Reading M-b gives 3.331 +/- 0.268 over 24 units, flight only, and is not judged. Leaving out the maps with cell_number -1 (a variant) the readings disagree in the same way (p 0.40 against ES, 0.020 against MC), with 21 FD9 maps, the paper's count. What the answers of 2026-10-20 changed. The criterion now runs each preflight epoch on its own under S-a and S-b, prefers neither, and keeps the claim blocked when they disagree, so the block that the attempt had applied pending the answer is now the criterion's own; the outcome is blocked before and after. New claim; 001754-knierim-2000-02 had none like it.

  How the test differed from the paper’s method

  • As for si-comparable-to-preflight (M-a activity by map rate).
• ⊘BlockedMain text, paragraph beginning "The rats occasionally turned around on the track", sentences 1 and 2· Methods underspecified

  Some cells had place fields only when the rat moved counterclockwise through a location, showing the direction selectivity seen on such tracks on Earth.

  Some cells demonstrated place fields when the rat was moving only counterclockwise through a single location, thus demonstrating the direction selectivity that is seen on such tracks under normal terrestrial conditions

  Main text, paragraph beginning "The rats occasionally turned around on the track", sentences 1 and 2

  The criterion needs, per bin, the time spent moving counterclockwise and clockwise. The dandiset gives position only at spike times (7.6 to 35.7 samples per second in the ES epochs, gaps up to 12.7 s) and one occupancy map per epoch with no direction, and nothing gives the track's layout in the image from which the direction along the circuit could be read. The claims file names no outside data. No entry of the zip files of the NASA record (OSD-968) has a position-like name, but its spike files (.cel) and the archives nested in its processed files were not opened, so whether it holds continuous position is not known. The answer of 2026-10-20 says the paper gives no method, no time base by direction and no track layout and allows no reading without them, and the criterion now blocks the claim as methods_underspecified in that case; before the answer the attempt blocked it as data_missing. Untested before and after. New claim.

• ⊘BlockedMain text, paragraph beginning "The rats occasionally turned around on the track", sentences 3 and 4· Data missing from the dataset

  Hippocampal EEG showed normal theta rhythm during active locomotion, in all rats on flight days 4 and 9.

  Hippocampal EEG activity was also ... recorded during baseline and behavioral sessions in all rats on FD4 and FD9. Although the small number of subjects precluded a statistical analysis, they showed normal theta rhythm during active locomotion

  Main text, paragraph beginning "The rats occasionally turned around on the track", sentences 3 and 4

  No file holds EEG or any continuous voltage (no ElectricalSeries or LFP in the listing of every file). The NASA record OSD-968 has no passage on EEG, theta, ripples or LFP, and no name inside its 26 zip files suggests EEG. Same outcome as in 001754-knierim-2000-02 (later removed from that record).

• ⊘BlockedMain text, paragraph beginning "The rats occasionally turned around on the track", sentence 4· Data missing from the dataset

  Hippocampal EEG showed normal sharp waves and ripples during quiet inactivity in the sleeping pouch, on both flight days.

  they showed normal theta rhythm during active locomotion and normal sharp waves and ripples15 during quiet inactivity in the sleeping pouch on both flight days.

  Main text, paragraph beginning "The rats occasionally turned around on the track", sentence 4

  No EEG in the dandiset, nor among the names of the OSD-968 files (see theta-normal). Same outcome as in 001754-knierim-2000-02 (later removed from that record).