Claims 26
From the paper Spatiotemporal distribution of location and object effects in reach-to-grasp kinematics · claims file · checked 2026-10-23: ready
- table2-trial-counts
In the three sessions the monkeys performed 768, 824 and 745 trials (L, X, Y), of which 717, 788 and 688 were successful and 685, 752 and 637 were analyzed.
during these sessions the monkeys performed from 745 to 824 total trials, of which from 688 to 798 were performed successfully, with from 637 to 752 trials being retained for analysis
Results, paragraph 1; Table 2How it will be judged
Counts per session from the trial records; analyzed trials are successful trials not preceded by an error trial (Methods, Behavioral Task). The text gives 798 as the largest success count and Table 2 gives 788 for monkey X; the table's 788 is used, and 798 is a second reading of that count. The exact counts are reported. Reproduced when all nine are within tolerance; partially reproduced when the analyzed counts are within tolerance and some other count is not.
- table2-movement-time
The median duration from movement onset to contact differs among the monkeys (255, 235 and 335 ms for L, X and Y; Kruskal-Wallis P < 0.001), giving 52, 48 and 68 normalized samples from M to C.
The duration of the epoch from movement onset to peripheral object contact, which we refer to as the reach-to-grasp epoch, differed among the three monkeys ... with medians ranging from 235 to 335 ms
Results, paragraph 1; Table 2How it will be judged
Movement time from onset (common methods, both readings) to contact, over the analyzed trials. The paper prints the bound as "P <10-3" with a garbled glyph; it is read as P < 0.001. Table 2's footnote does not say whether the median and quartiles are over analyzed or all successful trials; readings (a) analyzed trials, (b) all successful trials (the Fig. 5 legend speaks of "all successful trials"). Reproduced when, under one reading, the three medians are within tolerance and a Kruskal-Wallis test across monkeys gives p < 0.001. The sample counts follow from the medians at 200 Hz and are reported.
- fig2-trial-count-coax-45
Monkey L successfully reached to, grasped and manipulated the coaxial cylinder at the 45 degree location in 28 trials of its session.
Each frame shows data from 28 successfully performed trials in which monkey L reached to, grasped, and manipulated the coaxial cylinder at the 45° location
Figure 2 legend; Results, Temporal Trajectories of Individual Joint Angles, paragraph 1How it will be judged
Analyzed trials (common methods) of session L20120928 with that object and location. The legend says "successfully performed"; readings (a) analyzed trials, (b) all successful trials. Reproduced when either count is within tolerance.
- fig2-joint-patterns
In monkey L's movements to the coaxial cylinder at 45 degrees, the distal forearm speed has a bell-shaped profile, ShoulderIER moves monophasically from one position to another, and ElbowFlex first increases and then decreases.
although the speed of the distal forearm point showed the conventional bell-shaped profile, the trajectories of individual joint angles showed a variety of temporal patterns. Only ShoulderIER showed a monophasic transition from one angular position to another. The elbow flexion/extension (ElbowFlex) angle first increased and then decreased
Results, Temporal Trajectories of Individual Joint Angles, paragraph 2; Figure 2How it will be judged
The trials of fig2-trial-count-coax-45, time-normalized (common methods), the mean trajectory over trials from M to C. The paper gives no measure for these shapes, so they are read as follows (open, a reading of the words). Bell-shaped is a single interior maximum of the mean distal forearm speed; the distal forearm point is the point 1.1 cm deep to the five distal forearm markers (Methods), and the mean of the five markers is a variant. Monophasic is a mean ShoulderIER whose derivative keeps one sign from M to C. ElbowFlex is read as a mean whose maximum lies inside the epoch with a rise before it and a fall after it. Reproduced when all three hold; partially reproduced when two hold.
- fig2-hand-opening-two-phases
In the same movements, the ThumbFlex, four MCP, ThumbMCP, IndexPIP, MiddlePIP, RingPIP and ThumbAbAd angles extend (decrease) first, pause or slightly reverse near the middle of the epoch, near the time of peak transport speed, then extend again, and keep extending, opening the hand, until just before contact.
All of these angles initially decreased (extended), paused or slightly reversed near the middle of the reach-to-grasp epoch, and then decreased again, apparently separating the time course of hand opening into two phases ... all of these latter joints continued to extend (decreasing values), opening the hand, until immediately prior to peripheral object contact
Results, Temporal Trajectories of Individual Joint Angles, paragraphs 2 and 3; Figure 2How it will be judged
The trials of fig2-trial-count-coax-45, the mean time-normalized trajectory of each of the ten angles the paper lists (ThumbFlex, the four MCP, ThumbMCP, IndexPIP, MiddlePIP, RingPIP and ThumbAbAd; LittlePIP is not in this list, the paper puts it with the angles that decrease, increase and decrease again). No measure is given, so the words are read as follows (open). A pause is a local minimum of the extension speed (the negative of the mean angle's derivative), or a reversal to positive derivative, between 25% and 75% of normalized time; its time is compared with that of peak distal forearm speed. Extension until just before contact is the mean angle's minimum lying at or after 80% of normalized time. Reproduced when both hold for at least 8 of the 10 angles; partially reproduced for 5 to 7.
- fig4-armps-time-course
For ArmPS in monkey L, the location effect rises until about midway between onset and contact and then declines to small but still significant values before contact; the object effect stays small until shortly before the midpoint and then rises until contact, when it accounts for almost all the variance; the interaction is significant but small throughout.
After movement onset (M), the main effect of location (red) increased until approximately midway between movement onset and peripheral object contact (C) and thereafter decreased, reaching small but still significant values prior to contact. In contrast, the main effect of object (blue) remained small until shortly before the midpoint and then rose progressively until contact, at which time the object effect accounted for almost all of the variance in ArmPS angle
Figure 4; Results, Effects of Location and Object on Joint Angles as a Function of Time, paragraph 2How it will be judged
ANOVA effect sizes and significance as in the common methods, ArmPS, session L20120928, each sum of squares reading. "Approximately midway" and "almost all" are not given numbers; they are read as a location peak between 30% and 70% of normalized time, and an object eta2 at C of at least 0.9 (the default 10% tolerance on all of the variance, 1.0). Reproduced when, under every sum of squares reading, the location peak lies in that window, location eta2 at C is below its peak and significant, object eta2 at C is at least 0.9, object eta2 is below location eta2 at 30%, and the interaction eta2 is below both main effects at C and significant at some time from M to C. Partially reproduced when the order of the location peak and the object rise holds but a number does not. Readings that disagree make the claim blocked as methods_underspecified.
- fig5-location-proximal-object-distal
Across the 22 angles in all three monkeys, the largest location effects are in the shoulder and wrist angles and the largest object effects in the digit angles.
As expected, the largest effects of location were found in proximal joints, appearing as prominent red traces in shoulder and wrist angles, while the largest effects of object were found in the joints of the hand
Figure 5; Results, Effects of Location and Object on Joint Angles as a Function of Time, paragraph 3How it will be judged
ANOVA effect sizes from the common methods, each monkey. Reproduced when, in each monkey, the angle with the largest peak location eta2 from M to C is a shoulder or wrist angle (Table 1 groups Shoulder and Wrist), and the angle with the largest peak object eta2 is one of the 15 digit angles. The mean peak location eta2 over shoulder and wrist angles, and over digit angles, is reported, and likewise for object. Partially reproduced when this holds in two monkeys.
- fig5-early-significance-concurrent
For a given angle, location and object effects both typically become significant early in the movement, close to movement onset, and evolve concurrently, with small but significant interaction effects.
For any given joint angle, although location and object effects differed in magnitude, both typically attained significance early in the course of the present movements and evolved concurrently, with small but significant interaction effects (cyan traces) present as well
Results, Effects of Location and Object on Joint Angles as a Function of Time, paragraphs 3 and 5; Figure 5How it will be judged
ANOVA significance from the common methods. For each angle and monkey, the first normalized time at or after M where the location effect is significant, and the same for object and interaction. "Early" is not given a number; read as at or before 30% of normalized time, the paper's early sample point. Reproduced when, in each monkey, both location and object first become significant by 30% in more than half of the 22 angles ("typically"), and the interaction is significant at some time from M to C in more than half. Partially reproduced when this holds in two monkeys.
- fig5-shoulderier-location
ShoulderIER is dominated by location effects with little or no object effect.
ShoulderIER was dominated by location effects with little or no object effect
Results, Effects of Location and Object on Joint Angles as a Function of Time, paragraph 4; Figure 5How it will be judged
Qm of ShoulderIER (Eq. 2), each monkey. A Qm of 1 is a location effect with no object effect (Methods); the paper gives no bound for "little or no object effect", and it is read (a reading, not fixed by the paper) with the default 10% tolerance as Qm of at least 0.9. Reproduced when Qm is at least 0.9 at both the 30% and 90% points in all three monkeys, and partially reproduced when Qm exceeds 0.5 at both points in all three but is below 0.9 somewhere.
- fig5-shoulder-object-effects
ShoulderAbAd and ShoulderFlex both have clear object effects, in some cases beginning early in the reach.
The other shoulder angles, ShoulderAbAd and ShoulderFlex, both had clear object effects, however, in some cases beginning quite ... early in the reach-to-grasp
Results, Effects of Location and Object on Joint Angles as a Function of Time, paragraph 4; Figure 5How it will be judged
ANOVA significance from the common methods. "Quite early" is not given a number and is read as by 30% of normalized time, the paper's early sample point (a reading). Reproduced when the object effect of both angles is significant at some time from M to C in all three monkeys, and is significant by 30% of normalized time in at least one angle and monkey. Partially reproduced when the object effect is significant for both angles in two monkeys.
- fig5-elbowflex-object
ElbowFlex shows little location effect and a large object effect in each of the three monkeys.
the ElbowFlex angle showed surprisingly little location effect and instead had a large object effect in each of the three monkeys
Results, Effects of Location and Object on Joint Angles as a Function of Time, paragraph 4; Figure 5; compare Discussion, Effects of Object on Proximal Joint Angles, paragraph 1How it will be judged
ANOVA effect sizes from the common methods. The paper describes ElbowFlex in two ways, so two readings are listed. (a) The Results, object effect larger than location effect in each monkey, judged as mean object eta2 from M to C above mean location eta2 in all three. (b) The Discussion ("substantially larger than location effects in monkey L and similar in magnitude to location effects in monkeys X and Y"), judged as mean object eta2 above mean location eta2 in L, and the ratio of the two means between 0.5 and 2 in X and Y (the bounds of "similar" are a reading, not fixed by the paper). Fig. 6A's Qm of 0.55 and 0.54 at 90% in X and Y agrees with (b). Reproduced when either reading holds; the means and ratios are reported.
- fig5-wrist-comparable
The wrist angles ArmPS, WristRU and WristFlex each have location and object effects of comparable magnitude.
The wrist angles, ArmPS, WristRU, and WristFE, each had location and object effects of comparable magnitude
Results, Effects of Location and Object on Joint Angles as a Function of Time, paragraph 4; Figure 5How it will be judged
ANOVA effect sizes from the common methods; the text's WristFE is WristFlex. The paper gives no bound for "comparable"; it is read as a ratio of peak object eta2 to peak location eta2 from M to C between 0.5 and 2 (a reading, not fixed by the paper). Reproduced when this holds for the three wrist angles in all three monkeys; partially reproduced when it holds for at least 6 of the 9 angle-monkey pairs. The ratios are reported.
- fig5-hand-object-predominates
In all joint angles of the hand object effects predominate, with small location effects in some, particularly in monkey L; object effects in proximal joints are somewhat more marked than location effects in distal joints.
In all joint angles of the hand, object effects predominated. Nevertheless, small location effects were present in some, particularly in monkey L ... Overall, object effects in proximal joints were somewhat more marked than location effects in distal joints
Results, Effects of Location and Object on Joint Angles as a Function of Time, paragraph 4; Figure 5How it will be judged
ANOVA effect sizes from the common methods; hand angles are the 15 digit angles. The paper gives no measure; predominate is read as peak object eta2 above peak location eta2 from M to C (a reading). The paper uses "proximal" both for shoulder and wrist (paragraph 3) and for shoulder and elbow as against the wrist and digits (paragraph 5), so proximal angles have two readings, (a) the four shoulder and elbow angles, (b) these and the three wrist angles; distal angles are the 15 digit angles in both. Reproduced when this holds for all 15 digit angles in all three monkeys, the mean over digit angles of peak location eta2 is largest in monkey L, and the mean peak object eta2 of the proximal angles exceeds the mean peak location eta2 of the 15 digit angles, pooled over monkeys, under either reading of proximal. Partially reproduced when the first part holds for at least 40 of the 45 angle-monkey pairs.
- fig5-location-early-object-late
Location effects generally become large early in the movement and object effects later, typically peaking only in the second half; in the shoulder, location effects keep rising through the reach, while in the wrist and digits they rise, peak and decline as object effects increase.
location effects generally became large early in the course of these movements, while object effects generally became large later, typically reaching their maxima for a given joint only in the second half of the movements. In the shoulder joints, location effects typically continued to increase throughout the reach-to-grasp; but in the more distal joints of the wrist and digits, location effects typically rose initially, peaked, and then declined as the object effects increased
Results, Effects of Location and Object on Joint Angles as a Function of Time, paragraph 5; Figure 5How it will be judged
ANOVA effect sizes from the common methods. For each angle and monkey, the normalized times of the location and object peaks from M to C. "Typically" is read as more than half of the angles. Reproduced when, in each monkey, (1) the object peak lies after 50% for more than half of the 22 angles, (2) the object peak is later than the location peak for more than half of the 22 angles, (3) the location peak lies after 90% for at least two of the three shoulder angles, and (4) the location peak lies before 90% and location eta2 at C is below the peak, for more than half of the 18 wrist and digit angles. "Typically", the 90% bound for "continued to increase" and "peaked, and then declined" are readings, not fixed by the paper. Partially reproduced when (1) and (2) hold in all monkeys and (3) or (4) fails.
- fig5-interaction-small
Interaction effects, while often significant, are typically small; for no joint angle does the interaction effect account for more variance than either main effect.
Third, interaction effects, while often significant, typically were small. For no joint angle did the interaction effect account for more variance than either the location effect or the object effect
Results, Effects of Location and Object on Joint Angles as a Function of Time, paragraph 6; Figure 5How it will be judged
ANOVA effect sizes from the common methods. The sentence allows two readings; the Discussion adds that rarely was an interaction larger than both main effects. (a) For every angle and monkey, the peak interaction eta2 from M to C is below the larger of the two peak main effects. (b) For every angle and monkey, at no sample from M to C is the interaction eta2 above both main effects at that sample. Reproduced when either reading holds for all 66 angle-monkey pairs; partially reproduced when it holds for at least 60. The number of pairs and samples where the interaction exceeds one or both main effects is reported.
- fig6a-qm-30-90
At the 30% point Qm varies widely across angles, while at the 90% point only the three shoulder angles in each monkey, ElbowFlex in X and Y (0.55, 0.54) and WristRU in Y (0.56) have Qm above 0.5; all other angles have Qm below 0.5.
Figure 6A shows that, at the 30% time point, different joint angles had a wide variety of Qm ratios, but by the 90% time point only a few joint angles, particularly all three shoulder angles in each of the three monkeys, had Qm ratios ... 0.55 and 0.54, respectively) and WristRU in monkey Y (0.56). At the 90% time point, all other joints had Qm ratios
Figure 6A; Results, Effects of Location and Object on Joint Angles as a Function of Time, paragraph 7How it will be judged
Qm (Eq. 2) at the 30% and 90% points (common methods), each sum of squares reading. Reproduced when, at 90%, the angles with Qm above 0.5 are exactly the nine shoulder angle-monkey pairs plus ElbowFlex in X and Y and WristRU in Y, the three values are within tolerance, and at 30% each monkey has Qm values on both sides of 0.5. Partially reproduced when the nine shoulder pairs are above 0.5 and at most three of the other 57 pairs fall on the other side of 0.5 from the paper. Readings that disagree make the claim blocked as methods_underspecified.
- fig6b-qx-small
Qx is below 0.4 for the large majority of angles at both 30% and 90%; at 30% only WristFlex and ThumbFlex in monkey X and ThumbFlex in monkey Y exceed 0.4, none in monkey L, and at 90% all are below 0.4.
At the 30% time point, only monkey X showed Qx ratios ... in two joints angles, WristFlex and ThumbFlex, indicating substantial location ... object interactions early in the movement. In monkey Y only ThumbFlex had a large Qx. In monkey L no joint angles had a large Qx at the 30% time point. By the 90% time point, all Qx ratios were
Figure 6B; Results, Effects of Location and Object on Joint Angles as a Function of Time, paragraph 8How it will be judged
Qx (Eq. 3) at the 30% and 90% points (common methods), each sum of squares reading. Reproduced when the set of angle-monkey pairs with Qx above 0.4 at 30% is exactly WristFlex and ThumbFlex in X and ThumbFlex in Y, and no pair exceeds 0.4 at 90%. Partially reproduced when at most three of the 132 pair-time values fall on the other side of 0.4 from the paper. Readings that disagree make the claim blocked as methods_underspecified.
- fig7-mean-eta-30
Averaged over the 22 angles at the 30% point, location effects are as large as or larger than object effects (location/object/interaction, L 0.22/0.08/0.05, X 0.15/0.16/0.05, Y 0.12/0.11/0.04).
At the 30% time point, location effects were as large as, or somewhat larger than, object effects (location/object/ interaction: monkey L: 0.22/0.08/0.05; monkey X: 0.15/0.16/ 0.05; monkey Y: 0.12/0.11/0.04)
Figure 7; Results, Effects of Location and Object on Joint Angles as a Function of Time, paragraph 9How it will be judged
Mean over the 22 angles of eta2 (Eq. 1) for each effect at the 30% point, under every sum of squares reading and both placements of the maximum SSError (common methods). The tolerance is the default 10%, tight for two-decimal values; the values printed to two decimals are compared after rounding. The sum of squares type and the SSError placement are method details the paper leaves open, so the outcome is the one all combinations give. Reproduced when all nine values are within tolerance under every combination; partially reproduced when, under every combination, location is at least object minus 0.01 in every monkey and some value misses tolerance. Combinations that give different outcomes make the claim blocked as methods_underspecified. The values under every combination are shown.
- fig7-mean-eta-90
By the 90% point object effects are substantially larger on average than location effects in all three monkeys (location/object/interaction, L 0.12/0.52/0.07, X 0.13/0.44/0.04, Y 0.20/0.36/0.08).
By the 90% time point, still prior to contact with the object, object effects were substantially larger on average than location effects in all three monkeys (location/ object/interaction; monkey L: 0.12/0.52/0.07; monkey X: 0.13/ 0.44/0.04; monkey Y: 0.20/0.36/0.08)
Figure 7; Results, Effects of Location and Object on Joint Angles as a Function of Time, paragraph 9How it will be judged
As fig7-mean-eta-30, at the 90% point. Reproduced when all nine values are within tolerance under every combination of readings; partially reproduced when, under every combination, mean object eta2 exceeds mean location eta2 in all three monkeys and some value misses tolerance. Combinations that give different outcomes make the claim blocked as methods_underspecified.
- fig7-time-course
Averaged over angles, location and object effects begin to appear at about the same time; after the 30% point object effects grow more than location effects, which decline somewhat in L and X and stay about constant in Y; the proportions stay about constant after contact; interaction effects stay small.
Main effects of location and of object began to appear at approximately the same time on average ... Thereafter, object effects grew more than location effects, which actually declined somewhat in monkeys L and X, although remaining relatively constant in monkey Y ... The ratio of location, object, and interaction effects then remained relatively constant after peripheral object contact
Figure 7; Results, Effects of Location and Object on Joint Angles as a Function of Time, paragraph 9How it will be judged
Mean eta2 over the 22 angles as a function of normalized time (common methods). The paper gives no measure for these words; they are read as follows (open). Begin to appear is the first sample at or after M where the mean exceeds 0.05; the two times are compared, and "approximately the same" is a difference of at most 10% of normalized time. Growth from 30% to 90% is larger for object than location in each monkey. Location at 90% is below location at 30% in L and X, and not below it in Y (the paper's own values for Y rise from 0.12 to 0.20, so "relatively constant" is read as not declining). Constant after contact is Qm of the mean eta2 values staying within 0.1 of its value at C over the samples after C. The interaction mean is below both main effect means from 30% to C. Reproduced when all parts hold in all monkeys; partially reproduced when the growth part holds and another fails.
- fig8a-lda-location
Location given the object is best predicted by shoulder angles, which reach nearly 100% before contact, then by wrist angles, which match shoulder angles early but fall below them by the midpoint; digit angle groups exceed chance early, become less accurate than proximal groups shortly after onset, and still carry some location information just before contact.
In predicting location, shoulder joint angles (blue) were easily dissociated for different locations of a given object, reaching nearly 100% correct before peripheral object contact. Wrist angles (red) were next most strongly predictive of location. In the first part of the movements, wrist angles predicted location as accurately as shoulder angles in all three monkeys ... Nevertheless, note that just prior to peripheral object contact (C), wrist joints and even digit joints still had some degree of discriminable information on location
Figure 8A; Results, LDA of Joint Angles as a Function of Time, paragraph 2How it will be judged
LDA of location given object, from the common methods, each prior and interval reading. Nearly 100% is read with the default 10% tolerance as at least 90% correct; "the first part of the movements" as the 30% point and "by the temporal midpoint" as the 50% point (readings, not fixed by the paper). In each monkey, (1) shoulder accuracy at least 90% at some sample from 90% to C; (2) the mean accuracy over the samples from 50% to C orders shoulder above wrist, and wrist above each digit group; (3) the shoulder and wrist 95% intervals overlap at the 30% point and wrist upper bound is below shoulder accuracy at the 50% point; (4) each digit group's interval lower bound is above chance at the 30% point and at the 90% point; chance is the mean of 1/(locations available) over objects. Reproduced when all parts hold in all three monkeys under every reading; partially reproduced when (1) and (2) hold in all monkeys. Readings that disagree make the claim blocked as methods_underspecified.
- fig8b-lda-object
Object given the location is predicted similarly by all five angle groups, each reaching near 100% correct before contact in all three monkeys, including the shoulder angles.
In all three monkeys, all five joint-angle groups reached levels of object prediction close to 100% correct prior to peripheral object contact
Figure 8B; Results, LDA of Joint Angles as a Function of Time, paragraph 3How it will be judged
LDA of object given location, from the common methods, each prior and interval reading. Close to 100% is read with the default 10% tolerance as at least 90% correct. Reproduced when each of the five groups reaches at least 90% at some sample from 90% of normalized time to C, in all three monkeys, under every reading; partially reproduced when this holds for at least 12 of the 15 group-monkey pairs. Readings that disagree make the claim blocked as methods_underspecified.
- fig9-hand-configuration-location
At the 90% point in monkey L, index finger flexion varies modestly with location for the perpendicular and coaxial cylinders but little for the button and sphere, and hand configuration varies more with object than with location.
Modest variation in the degree of index finger flexion is evident depending on location for the perpendicular and coaxial cylinders, but relatively little if any variation in index finger flexion was present for the button or sphere ... A modest degree of locationdependent variation in hand configuration again can be appreciated, although more variation in hand configuration depended on object
Figure 9; Results, Hand Configurations and Reach Trajectories, paragraphs 1 and 2How it will be judged
As in Fig. 9 (legend and Results), session L20120928, the 90% point, marker positions averaged over successful trials of each location-object combination, translated and rotated to co-register the five hand markers. The paper gives no measure of spread; the following is a reading (open). Location spread per object is the mean distance of the four index finger markers' combination means from their mean over locations; object spread per location likewise over objects, using all 20 digit markers, compared with location spread using all 20 digit markers. Reproduced when index location spread is larger for each cylinder than for each of button and sphere, and mean object spread exceeds mean location spread. Partially reproduced when one of the two holds. If the dataset holds joint angles and not markers, the across-location SD of mean IndexMCP and IndexPIP per object is used instead, as a variant.
- fig10-forearm-trajectory-object
Distal forearm point trajectories vary mainly with location, but at a given location they also depend on the object from early in the reach, with little overlap between objects by the end; at contact the point is typically highest for the sphere, intermediate for the coaxial cylinder and lowest for the button.
At any given location, however, the trajectory of the distal forearm point through space depended to some extent on the object to which the hand was being transported, beginning quite early in the reach ... by the time of peripheral object contact, the distal forearm point typically was positioned highest for the sphere (orange), intermediate for the coaxial cylinder (cyan), and lowest for the button (violet) at any of the locations sampled
Figure 10; Results, Hand Configurations and Reach Trajectories, paragraphs 3 and 4; Discussion, Effects of Object on Proximal Joint Angles, paragraph 1How it will be judged
Distal forearm point from the distal forearm markers in earth-fixed coordinates (as the Results say), successful trials, normalized time. Fig. 10 does not name the monkey; readings (a) monkey L (as Figs. 3 and 9), (b) each monkey. The paper gives no measure here (its quantitative support is Fig. 8B); the following is a reading (open). The LDA of object given location of the common methods, run on the point's three coordinates; reproduced when its accuracy interval lower bound exceeds chance by the 30% point and accuracy reaches at least 90% at C, and the mean up coordinate at C is ordered sphere above coaxial cylinder above button at more than half of the locations where all three were presented. Partially reproduced when one of these holds. Under reading (b) the result per monkey is reported and the claim judged on L.
- elbow-point-trajectory-object
The elbow point trajectory, set by shoulder rotation alone, also varies mainly with location and to some extent with object from early in the reach.
These plots likewise demonstrated variation depending primarily on location, but also depending to some extent on the object, beginning early in the reach
Results, Hand Configurations and Reach Trajectories, paragraph 4 (plots not illustrated)How it will be judged
The elbow point estimated as in the Methods (from the three upper arm and three proximal forearm markers, 1.7 cm deep), the same reading and measure as fig10-forearm-trajectory-object, without the vertical ordering, and with a location LDA given object on the same coordinates. "Primarily on location" is read (a reading, the paper gives no measure) as the comparison of accuracies above chance below. Reproduced when object classification given location exceeds chance by the 30% point and, at C, location accuracy given object minus its chance level exceeds object accuracy given location minus its chance level. The monkey is not named; judged on L with the others reported.
- fig8-hand-shape-invariant-to-location
Late in the reach, location-dependent differences in hand shape are minor compared with object-dependent differences in distal forearm position, as the lower accuracy of predicting location from digit angles compared with the high accuracy of predicting object from shoulder angles confirms; grasp shape is comparatively invariant to location.
The late location-dependent differences in hand shape (Fig. 9) overall seemed relatively minor compared with the objectdependent differences in distal forearm position (Fig. 10). This impression was confirmed by the lower prediction accuracies of location from digit joint angles compared with the high prediction accuracy of object from shoulder joint angles (Fig. 8)
Discussion, Effects of Location on Distal Joint Angles, paragraph 2; Figure 8; AbstractHow it will be judged
LDA from the common methods, each prior and interval reading. The paper compares percent correct and does not name the time; "late" is read as the 90% point (the time of Fig. 9) and as C, two readings of the text. Reproduced when, in each monkey, the object accuracy given location from shoulder angles exceeds the location accuracy given object from each of the three digit groups, at the 90% point or at C; partially reproduced when this holds in two monkeys. Accuracy above chance for each is reported as well. Prior and interval readings that disagree make the claim blocked as methods_underspecified.
Pipeline
- Paper
- done (4 of 4)
- Claims
- done
- Claims check
- done (ready)
- Attempt
- todo
- Questions
- none
- Verified
- none