MSSPL Archive

About the study

The Marsh Street Sleep and Perception Laboratory (MSSPL) was a research unit of the Department of Neurobehavioral Medicine at Halloran-Reeve University. This is a volunteer mirror of what remains.

Between October 2004 and April 2008 the laboratory ran the Longitudinal Study of Convergent Nocturnal Imagery (LSCNI). Eighty-four adult volunteers, in seven cohorts (A to G), each spent twenty nights in one of four sleeping rooms. Overnight polysomnography was recorded. A wake prompt was delivered at the first REM onset after 02:00, and the participant completed a handwritten dream log within five minutes. The logs were scanned at 08:00 and the originals sealed.

The participants had not been told what to describe. The form said only "Describe what you saw." Even so, 61 of the 84 described the same building, in similar order and in similar terms, and had not been in contact with one another. The study team calls it the Annex. That is a participant term. The published papers say "building."

The laboratory was archived in January 2010. The papers, personnel files, documents and logs below are transcribed as found, including errors, corrections and withheld pages. No mechanism is proposed anywhere in the archive, and none is proposed here.

Open the internal field guide: Annex Architecture, a working description

Internal field guide and glossary · Working draft, Rev. 11

Marsh Street Sleep and Perception Laboratory. Internal field guide and glossary.
Prepared by E. Hale, B. Okonkwo-Strand and A. Fitch, from the LSCNI logs (LOG-0001 to LOG-1680, plus the pilot file #oyelaran).
Document status: Working draft, Rev. 11. Supersedes Rev. 10. (Rev. 9 was never issued; the register entry reads "not found.")
Distribution: study team only. Not for circulation to participants.

0. Purpose and caution

This guide collects what participants have consistently reported about a shared environment that the study team calls, for convenience, the Annex. The name was supplied by participants (first in the pilot file, February 2004, and in the main series in LOG-0577, S-029: "I was in the Annex again"). The word "building" is used in the published papers; the word "Annex" is used here because it is what the participants say.

The description is assembled from 61 of 84 participants who reported the building, across 20 nights each. Of the 61, 19 also described a figure with a headwear stimulus (the Chairman, Section 2.5); those 19 are included in the 61, not added to it. The other 42 reported the building only, and 23 of the 84 did not report it. It is a composite. No participant has described all of it. Where participants disagree, the disagreement is stated. Where they do not disagree, the agreement is stated, and the reader is advised that nothing in the protocol would have produced it.

The guide does not propose a mechanism. It does not propose a purpose. It does not propose that the Annex is anywhere.

Study staff are asked to refer to any figure with an item of head covering as the Chairman, or as a headwear stimulus, in writing and in speech. This is a convention of the laboratory. No reason for the convention has been recorded.

1. Layout

1.1 The waiting room

The waiting room is the first place reported. Twenty-two of the first 36 participants described it as the opening setting; all 61 reporters eventually did.

Seating. Rows of chairs, joined at the legs, in a color participants variously call "green," "grey-green" or "the color of a hospital." The seats are warm. Participants describe sitting without having decided to sit.

Slips. A small paper slip with a number. Participants do not recall taking it. Several say they already had it.

The board. A display of numbers, changing slowly. No participant has described the board reaching their number. Several have described it reaching the number before theirs, and then going blank.

Light. Even, flat, without a visible source. No shadows are reported on the floor, though participants report shadows on the walls.

Sound. A low hum, described as "a refrigerator in the next room." The hum is present in 53 of 61 accounts. In six it is described as rising when the board changes.

The clock. Present in 14 of the first 22 accounts. Where the time is legible, it is most often 03:07.

The room is occupied. Participants consistently report that the waiting room is already in use: a coat on a chair, a cup still warm, a person who has "just stepped out." The occupants are never seen.

1.2 The counters

A row of counters along one wall, numbered, with a window at each. Participants report between six and twelve. Each is staffed by a Clerk. The counters are used for: receiving slips, handing back items, and "showing." (See Section 2.)

Participants who approach a counter report that the Clerk "had already written something." What is written is not shown to them. Participants report being handed back something that belongs to them: a coat, a photograph, a key. In 13 accounts, the object returned was recognized as the participant's own; in five, it was recognized as belonging to someone the participant had not seen in some years. In none was the object found on waking.

1.3 The ledger rooms

Rooms behind the counters, reached by a door at the side of the waiting room. Narrow shelves, bound volumes with numbered spines, a long table with green-shaded lamps. Participants report that the volumes contain "everything" or "the records," and that they were unable to read them. Six participants reported finding a volume with a number matching a number they had been given. None was able to open it.

Several participants report that the ledger rooms smell of paper and "something like a school." The smell is the most reliably reported sensory detail after the hum.

1.4 The corridors

Long, carpeted, lit by a series of lamps at even intervals. Doors on both sides, numbered. The numbers do not repeat. Participants who attempt to count the doors report different totals each night, but agree that "the numbers go up and then they go up again."

The corridor floor is quiet. Participants report that footsteps are muffled and that they cannot hear their own. They can hear the Attendants.

Participants report that the corridor "continues past the point where it should have ended," and that turning around does not return to the waiting room unless an Usher is present.

1.5 The stair

The staircase is reported by 31 of the first 40 reporters. It is described as wide, wooden, with a handrail on the left going down. There is a landing midway in most accounts.

The Count. Going up, the participant counts a number of steps. Going down, they count one more. The number differs between participants, between 11 and 18 going up, and is stable within each participant. The difference is always one.

The Count has been described by participants as "not wrong, just how it is."

The light diminishes on the way down. Participants do not report fear. They report that the descent is "orderly."

1.6 Floor B

Reported by nine participants out of 31 who described the stair. In every account, the stair ends at a door marked "B." Six of the nine say they did not go through. Three decline to elaborate.

Participants who have described Floor B do so in the present tense, even when writing the log the next morning. One wrote: "It is warm and the lamps are lit and there are many of us." (LOG-0958, S-048, night 18.)

The team has no further information about Floor B. Any account that includes a description of what is on Floor B should be flagged to the principal investigator.

2. The Attendants

The Attendants are the staff of the Annex. Four ranks are recognized on behavioral grounds. None has been reported to run.

2.1 Ushers

The most commonly reported. They stand in the waiting room and in the corridors. They gesture forward with an open hand. They speak short phrases, mostly numbers, sometimes "this way." They do not touch participants. They are described as courteous. Several participants mention a small badge or a pin, which they could not read.

Ushers are reported to count participants in. The count is audible to participants and is usually their own number.

2.2 Clerks

Behind the counters, and in the ledger rooms. They write. They hand things across. They "show": a photograph, a face on a card, a number, a door. What is shown is nearly always something the participant has seen before, in waking life. In a minority of accounts it is something the participant has not yet seen. In one account (LOG-0417, S-021) the Clerk handed over a photograph of a kitchen the participant did not recognize. Six weeks later the participant moved to a house with that kitchen.

Note by the principal investigator: this account has not been verified. The participant's address history was not collected, and the study team does not hold it.

2.3 Wardens

Reported least often. They stand at doors. Participants report that doors "do not open" while a Warden is present beside them, and that the room "stays as it is." Wardens are rarely described in detail. They are reported to be "still."

The study team has not examined Wardens in relation to any condition, state or syndrome, and the reader should not do so.

2.4 Surveyors

Reported by eleven of the first 40. They measure rooms, walls, floors, doors. They carry a folding rule or a length of cord. They speak numbers aloud, usually in whole centimeters. In four of the eleven accounts, the participant recognized the room measured as their own bedroom.

Surveyors are the Attendants associated with carryover (see Glossary).

2.5 The Chairman

Not an Attendant. Reported by 19 participants, always in a reflective surface, never directly. The description differs in every report. Attendants are reported to defer to him ("they check with him"). No participant has reported hearing him speak.

Staff are asked not to discuss the Chairman in the corridor outside Room 3.

3. Observed regularities

The following six regularities have been observed in the logs. They are presented without interpretation.

3.1 Attendants do not run. No Attendant of any rank has been described moving faster than a walking pace (0 of 61 reporters). Seventeen participants described a figure running or moving quickly. In all seventeen, the figure was not an Attendant. The study has no classification for these figures. Coding rule 7 (Rev. 10) states: Do not code a running figure as an Attendant.

3.2 The Count. One more stair going down than going up (31 of 31). See 1.5.

3.3 The slip. Participants who report being asked to sign a slip, a form or a ledger have, in each case, an earlier log entry in their own record that predates the reported signing. In other words, they report being asked to sign something they appear to have already signed. This regularity is recorded in 11 accounts (LOG-0561 is typical). The study team does not hold a record of what was signed.

3.4 The reflection. The Chairman is seen only in reflective surfaces (19 of 19). Participants who have attempted to look at him directly report that "he is where he was, in the glass."

3.5 Displacement. Objects in the sleeping room are displaced by 2-7 cm on waking. No event has exceeded 7 cm (14 of 14 measured). Displacement is reported only by participants who have completed 18 or more logs.

3.6 The correction. Participants who keep their logs faithfully receive a correction. The correction takes the form of a second-hand annotation. It is brief. It is usually correct. (LOG-0752 is typical.)

4. Glossary of coined terms

Annex. The shared environment described above. Participant term.

Attendant. Any of the four ranks of staff described in Section 2.

Carryover. Any effect that appears to cross from the reported environment into the participant's waking surroundings. The principal example is displacement. Carryover is attributed to Surveyors on the basis of participant reports. This attribution is provisional.

Chairman. A figure with a headwear stimulus, seen in reflection. Participant term. The Chairman's relationship to the Attendants is hierarchical on the basis of 8 accounts.

Convergence. Agreement between the accounts of participants who have not been in contact. Measured by the convergence index (CI). Pooled CI across Cohorts A-G is 0.71.

The Count. The asymmetry in stair steps: +1 on descent.

Displacement. A change in the position of an object in a participant's sleeping room, observed on waking, of 2-7 cm.

Filing. The state of having been entered into the records of the Annex. Participants report being "filed" in the sense that a slip, a number or a ledger entry exists for them. Filing is not an event the study can observe. It is inferred from accounts of signing (3.3).

Headwear stimulus (HWS). Any item of head covering described on a figure in a log. Used in place of other terms by convention.

Log closure. The practice of ending a dream log with the single line "Done." The effects of closure are the subject of a manuscript that has not been issued (Paper 12).

Reciprocal annotation. The appearance of second-hand marginal notes in sealed logs. See Paper 5 (withdrawn).

5. What participants report on waking

This section is limited to what participants wrote. The study has not assessed these reports clinically and proposes no meaning for them.

Participants describe waking slightly rearranged. The most common phrase (22 accounts) is "like I was put back." Others: "the room is the right room but it was visited." A small number describe a short interval, of a few seconds to a minute, in which the room "felt measured." They describe familiar objects as briefly more specific: a chair that is clearly this chair.

Eleven participants report waking at the same moment, usually 03:07, and then, in the days after, finding that they woke at that time without a prompt. They do not report distress. They report a sense of having arrived.

Twelve participants wrote that they "knew a number" on waking: a stair count, a counter, a figure in centimeters. In some cases the number was later found to match a measurement of the room. This has been checked in three cases. We have not published it.

Participants do not report being harmed. Participants do not report being told anything. Several have written that they were shown something, and that on waking they could not say what.

The study team does not offer any advice to participants about sleep, wellbeing or health. Questions of that kind were referred to the participant's own physician, in accordance with the consent form.

6. Handling notes

Logs are to be kept in ink, per Form 7-C Rev. 6.
Annotations in a second hand are to be recorded on Form 7-F and not removed.
No member of staff is to write in a participant's log.
The log of any participant who has completed 18 or more entries is to be photographed on the day it is received.
Room 3 is to be kept locked when unoccupied.
The register entry for Rev. 9 is to remain as it stands.

Revision history: Rev. 8 (2005); Rev. 9 [not found]; Rev. 10 (2007); Rev. 11 (2009). Last modified 03:07.

Findings at a glance

Figures taken from the papers, the logs and the field guide. They are reported without interpretation.

84 participants enrolled, Cohorts A to G
79 consent forms on file (Form 4-C, v.1 to v.4)
61 reported the building; 23 of the 84 did not
19 of those 61 described a figure seen only in reflection: 19 of 19 accounts (Paper 10)
0.71 pooled convergence index, Cohorts A to G (control task: 0.18)
+1 stair difference exactly one: 31 of 31 reporters counted one more step going down than going up (Paper 2)
03:05-03:09 awakenings cluster here: 47 of 161 spontaneous terminal awakenings, where about 4.5 were expected (Paper 8)
1,680 numbered logs, of which 1,642 originals were located

The Count: steps going down / steps going up

Participants counted between 11 and 18 steps up and 12 and 19 down. The difference was always one.

Publications

PAPER 1

Published 2006

The Waiting Room Phenomenon

Hale E, Prasad I, Okonkwo-Strand B. Quarterly Review of Nocturnal Imagery 3(2): 41-58. 2006.

Participants in a longitudinal sleep study (LSCNI, n = 36 across Cohorts A-C) completed a structured dream log within five minutes of each wake prompt. Of the 36, twenty-two (61%) independently described the same initial setting: a waiting room with fixed seating, numbered slips, and a display board. Pairwise feature overlap between participants who had no contact was high (convergence index 0.64) relative to a control task in which twenty volunteers were asked to imagine a waiting room (0.18). Participants described the room as already occupied. A pilot participant (S-0, file #oyelaran) described the same room, in the same order, before the study protocol had been finalized. The phenomenon is reported descriptively. No mechanism is proposed.

Open full paper

Archive copy · Paper 1

Received 02 February 2006. Accepted 19 April 2006.

1. Introduction

Reports of recurring settings in nocturnal imagery are common and have been documented in survey form (Aldous & Penhallow, 1998). Such settings are typically idiosyncratic: the childhood house, the school corridor, the workplace. Cross-participant agreement on a setting is not expected, and where it has been reported, it has been attributed to shared cultural reference (Thorne-Vasquez, 2001).

During pilot work in 2004, a night nurse (S-0) described a waiting room containing "rows of chairs bolted together," a numbered slip, and a "board that never reached her number." Eight months later, an unrelated volunteer (S-006), who had been told only that the study concerned "recurring night imagery," described the same room. The wording of the second account was not prompted. This observation motivated the present analysis.

2. Methods

Participants. Thirty-six adults (Cohorts A, B, C; Subjects S-001 to S-036), recruited by flyer and screened by Dr. Prasad for primary sleep disorders (exclusions: seven). Consent was obtained on Form 4-C (v.1 for Cohorts A and B, v.2 for Cohort C) under protocol HSRC-04-118.

Design. Each participant completed twenty nights in the Marsh Street laboratory (Rooms 1-4) over a period of up to ten weeks. Overnight polysomnography (EEG, EOG, chin EMG) was recorded. A wake prompt was delivered at the first REM onset after 02:00. Participants completed the dream log (Form 7-C) by hand within five minutes. Logs were numbered sequentially (LOG-0001 onward), scanned by the study coordinator at 08:00, and the originals sealed.

Coding. Two coders (Fitch, Prasad) extracted setting features from each log using a 31-item checklist. Coders were blind to participant identity and cohort. In this paper, any figure described as wearing an item of head covering is recorded under a single code and is not analyzed further.

Convergence index. For each pair of participants reporting a building, the proportion of checklist features shared was computed (Wren & Prasad, in preparation, describes the index in full). Control: twenty volunteers (not sleep-study participants) were asked to describe an "imagined waiting room."

3. Results

Twenty-two of 36 participants (A: 7/12; B: 8/12; C: 7/12) reported a building across at least three nights. In all twenty-two, the waiting room was the first location described on the first night the building was reported.

Table 1. Features of the waiting room (n = 22 reporters)

Feature	Reporters	%
Rows of fixed seating	22	100
Door to corridor	20	91
Numbered slip or ticket	19	86
Display board with numbers	21	95
Lighting with no visible source	18	82
Reflective glass or polished surface noted	15	68
Clock present	14	64
Counter or window with a person behind it	17	77

Among the fourteen participants who noted a clock, nine reported a legible time. Six of the nine reported 03:07. The remaining three reported different times (02:41, 04:19, 03:55).

Table 2. Convergence index by cohort

Group	Pairs	Mean overlap
Cohort A	21	0.61
Cohort B	28	0.66
Cohort C	21	0.65
Pooled	70	0.64
Control (imagined room)	190	0.18

Participants were not asked about the waiting room; the log form contained only the prompt "Describe what you saw." No cohort differed significantly from the others (Kruskal-Wallis, p = 0.71). Twelve of the 22 reporters noted without prompting that the room "seemed already in use."

Pilot participant S-0 (file #oyelaran) is excluded from all tables. Her pilot account is summarized in Section 4.

4. Discussion

The agreement observed is greater than that expected from shared cultural reference. The design cannot exclude a common stimulus in the laboratory environment, though the four rooms differ in layout and no waiting-room-like feature was in use in the laboratory (the building has one storey and no reception area; a bay of six chairs outside Room 1 was not used by participants). A suggestibility effect via the log form was considered; Cohort A completed logs on blank paper, and Cohort A's convergence index (0.61) did not differ from Cohorts B and C.

The "already in use" observation is reported without interpretation. Participants who described it used similar phrases: "someone had just left," "the seats were warm," "my number was already on the board." One participant (S-014) wrote: "I did not take a slip. I had one."

The convergence index for the control task (0.18) indicates that overlap in imagined waiting rooms is low, which is the expected result. The study is limited by sample size, by the self-report nature of the logs, and by the absence of an independent record of the setting. A correction in data handling after Cohort A (a change from pencil to ink on Form 7-C, Rev. 6) is noted in the supplement.

Further work will examine what participants do in the waiting room (Section 5 of the protocol, Rev. 8).

Acknowledgments

Supported by the Ardent Foundation for Cognitive Research (AF-0412) and the Tindall Family Trust. We thank P. Pell-Kerrigan for technical support and the participants for their patience.

References

Aldous R, Penhallow M. (1998). Recurring settings in nocturnal reports: a descriptive survey. Quarterly Review of Nocturnal Imagery 1(1): 5-29.
Thorne-Vasquez L. (2001). Shared imagery and cultural reference: a reassessment. Archives of Observational Dream Studies 3(2): 88-104.
Wren T, Prasad I. (in preparation). Convergence index: construction and validity.
Hale E, Okonkwo-Strand B. (2005). LSCNI protocol, Rev. 8. Marsh Street Sleep and Perception Laboratory (internal).
Marsh Street Laboratory (2004-2005). Form 4-C, v.1 and v.2: Participant consent.

Reviewer comments (excerpt, Reviewer 2). "The agreement is striking. I would ask the authors to consider whether the flyer or the consent form could have supplied the vocabulary. I reviewed Form 4-C and do not think so. I would, however, ask the authors to explain the sentence in Section 3, paragraph 4: someone had just left. Who is meant?" — Author response: "This is a quotation from the participants. We have no further information."

PAPER 2

Published 2007

Stair Count Asymmetry

Wren T, Hale E. Midlands Proceedings in Sleep Research 14: 112-127. 2007.

Forty of 60 participants (Cohorts A-E) reported a building across multiple nights; thirty-one of the forty described a staircase with a countable number of steps. In every one of the thirty-one cases (100%), the participant counted exactly one more step on the way down than on the way up. Counts varied between participants (ascent 11-18, descent 12-19) but were stable within participants across nights (27 of 31 reported the identical ascent count every night). A control group of 24 volunteers asked to imagine a staircase produced a mean ascent-descent difference of 0.1 (SD 1.9). The asymmetry is reported as an observed regularity, designated "the Count." No explanation is offered.

Open full paper

Archive copy · Paper 2

Received 11 June 2007. Accepted 30 August 2007.

1. Introduction

Counting is among the most frequently recorded actions in nocturnal reports of stairs (Marchetti-Owusu, 1999). Participants in the Longitudinal Study of Convergent Nocturnal Imagery (LSCNI) had not been asked about stairs. The dream log contained no prompt referring to them. Following the observation of a waiting-room setting (Hale et al., 2006), the coders recorded all architectural features. A staircase was noted in the early logs of Cohort B, and a participant (S-017) wrote: "Fourteen up. Fifteen down. I counted twice."

2. Methods

Participants. Sixty adults (Cohorts A-E, S-001 to S-060). Forty reported a building. Thirty-one reported a staircase with step counts recorded in at least one log.

Extraction. Step counts were extracted from the logs by the first author (Wren) and independently by Mr. Fitch. Disagreements: none. Where a participant gave a count in words ("fourteen"), it was converted to a number. Where a participant gave a range ("about fifteen"), the log was excluded from the paired analysis.

Control. Twenty-four volunteers (staff of an unrelated administrative office) were asked to imagine climbing a staircase to a higher floor and returning, and to count the steps in each direction.

Analysis. The ascent-descent difference was computed for each participant, using the first log in which both counts appeared. Within-participant stability was computed across all logs with both counts.

3. Results

Table 1. Ascent and descent counts (n = 31)

	Mean	SD	Range
Ascent	14.2	2.1	11-18
Descent	15.2	2.1	12-19
Difference	1.0	0.0	1-1

Thirty-one of thirty-one participants (100%) recorded a difference of exactly +1 (descent minus ascent). In the control group, the mean difference was 0.1 (SD 1.9; range -4 to +4); eleven controls reported equal counts.

Table 2. Stability and associated features

Measure	Value
Identical ascent count on all nights reported	27/31
Ascent count changed by one or more between nights	4/31
Landing midway reported	23/31
Handrail on the left when descending	19/31
Light decreasing on descent	26/31
Stair described as "ending" at "B"	9/31

In the four participants whose ascent counts varied across nights, the difference remained +1 on every night. For example, S-026 recorded 13 up and 14 down on one night, and 15 up and 16 down on another.

Nine participants wrote that the staircase ended at a door or a sign labeled "B." Six of these nine added that they "did not go through." Three did not elaborate.

Of the 31 participants, nine recorded a time at the end of the staircase event (a clock, a watch, a display). The time was 03:07 in four cases.

4. Discussion

The paired difference is the most consistent quantitative finding of the LSCNI to date. The probability of 31 consecutive +1 differences under the control distribution is, by direct calculation, smaller than the precision of the available tables. We therefore report the result descriptively and do not attach a p-value.

Three artefacts were considered. First, transcription error: the raw logs were checked against the scans by a second reader (Okonkwo-Strand), with no discrepancies. Second, shared coding bias: the extraction was repeated blind by a coder who had not read the earlier papers; the result was the same. Third, physical staircases in the laboratory: the Marsh Street building has one storey and no stairs; the nearest staircase is in the adjoining Halloran Annex, which no participant entered.

The between-participant variation (ascent 11-18) suggests that the staircase is not a fixed object. The within-participant stability suggests that for each participant it is a fixed object. These observations are not obviously compatible. We note only that the +1 appears to be a property of the direction, not of the staircase.

Limitations. Step counting in a nocturnal report depends on the participant's attention at the time. Participants did not know that others had counted. Counts were obtained after waking.

A footnote: the first author counts steps aloud when walking. This habit was noted by colleagues before the analysis was begun, and was not a source of data.

Acknowledgments

Supported by the Ardent Foundation (AF-0412) and the Tindall Family Trust. We thank B. Okonkwo-Strand for log maintenance and P. Pell-Kerrigan for the scanned archive.

References

Marchetti-Owusu D. (1999). Counting in nocturnal reports. Archives of Observational Dream Studies 1(3): 41-55.
Hale E, Prasad I, Okonkwo-Strand B. (2006). The Waiting Room Phenomenon. Quarterly Review of Nocturnal Imagery 3(2): 41-58.
Wren T, Prasad I. (2007). Convergence Index: construction and validity. Annals of Sleep Cohort Methods 5(1): 33-49.
Prasad I, Marsh-Teague J. (2007). REM fragmentation across the 20-night protocol. Midlands Proceedings in Sleep Research 15: 9-24.
LSCNI protocol, Rev. 8 (superseded by Rev. 10). Internal.

Editor's note. The two reviewers' reports are available on request. Reviewer 1 wrote: "I checked the counts myself against the scans. I then checked them again." Reviewer 1 declined to review a revised version. The decision to publish the paper without a p-value was taken by the editor.

PAPER 3

Published 2007

Attendant Roles in Reported Nocturnal Figures

Haddad-Lindqvist N, Fitch A, Hale E. Archives of Observational Dream Studies 9(4): 203-226. 2007.

Among 40 participants (Cohorts A-E) who reported a building, 38 described one or more persons present in the building who were not the participant. We propose that these persons can be classified into four recurring roles, designated Ushers, Clerks, Wardens and Surveyors, on the basis of consistent behavior across independent accounts (inter-rater agreement κ = 0.81). Eleven participants also described a figure with a headwear stimulus; this figure was seen only in a reflective surface in all eleven cases and described differently in every report (κ = 0.04). No participant described any of the four roles moving faster than a walking pace. The classification is behavioral. No claim is made as to the identity or intent of the figures.

Open full paper

Archive copy · Paper 3

Received 20 July 2007. Accepted 15 October 2007.

1. Introduction

Persons in nocturnal imagery are usually familiar: relatives, colleagues, strangers drawn from daily life. The accounts collected in the LSCNI diverged from this pattern. Participants described people who "worked there," who were consistent in manner across nights, and who were not recognizable. Several used the word "staff." The terms Usher, Clerk, Warden and Surveyor were adopted from participants' own language: all four appeared spontaneously in the logs of at least two participants before any member of the research team used them.

2. Methods

Material. All logs from Cohorts A-E (n = 1,200 logs, 40 reporting participants) were screened for accounts of other persons.

Coding. Two coders (Fitch, Prasad) classified each described person by (i) location, (ii) action, (iii) interaction with the participant, and (iv) any item carried. A role was recognized when the same combination of location, action and carried item appeared in accounts from at least three independent participants. Coder agreement was assessed by Cohen's kappa.

Chairman reports. Participants used the term Chairman for the figure described below in 7 of 11 reports. In the analysis, any figure with an item of head covering is coded as HWS.

3. Results

Table 1. Role frequency (n = 38 describing other persons)

Role	Reporters	Typical location	Typical action
Usher	34	Waiting room, corridors	Gestures "forward"; speaks numbers or short directions
Clerk	28	Counters, ledger rooms	Writes in a ledger; hands objects back to the participant
Warden	15	Doors, thresholds	Stands beside a door; participants report doors "do not open" while present
Surveyor	11	Any room	Measures walls or floors with a folding rule or cord

Ushers did not touch participants in any account. Clerks were reported to "show" things (a face, a photograph, a number written on a card); thirteen participants described being shown an item they recognized from their own life. Wardens were seldom described in detail; they were reported to "keep a door." Surveyors were reported to measure rooms, and in four of the eleven reports the participant recognized the room measured as their own bedroom.

All four roles were described as walking. Seventeen participants described a figure moving quickly or running; in all seventeen the figure was not coded as any of the four roles (coding rule 7 was added to the protocol, Rev. 10). We have no further account of these figures.

Chairman reports. Eleven participants reported a figure with an HWS. In all eleven, the figure was seen in a reflective surface: a polished counter, a window, the glass of a framed notice. None described the figure by direct view. Four participants provided reports from more than one night; none of the four gave the same description twice. Descriptions of height ranged from "short" to "taller than the doorway"; descriptions of build ranged from "thin" to "heavy"; one participant (S-031) wrote that the figure "looked like me." Inter-rater agreement for physical description was κ = 0.04.

Two participants reported that Ushers and Clerks "asked" the figure before acting ("they check with him"). No participant reported hearing the figure speak.

4. Discussion

The role structure is the clearest evidence in the data for a shared environment of some kind. The roles are distinguished by function rather than appearance, which is unusual for imagery of this type: appearance is typically what varies, and function is typically what is lost.

The Chairman reports are reported separately because they differ in kind. The figure is never seen directly; there is no consistent description; and participants refer to the figure with a tone the coders recorded as "deferential" in eight of eleven reports. The figure's role in relation to the Attendants appears hierarchical, though no participant described the figure giving a direction.

We considered whether the four-role scheme was imposed by the coders. The scheme was derived from participants' language and tested on a held-out set of logs (Cohort E), where it classified 94% of described persons. We also considered the possibility that participants learned the scheme from one another; participants in the same cohort were not in contact during the study and slept in separate rooms.

Limitations

The classification depends on logs. Two reviewers suggested alternative accounts for the observed regularities; we examined each against the measures collected in the study and found none supported. We do not rule them out.

Acknowledgments

Supported by the Ardent Foundation (AF-0412) and the Tindall Family Trust. Data management: B. Okonkwo-Strand.

References

Hale E, Prasad I, Okonkwo-Strand B. (2006). The Waiting Room Phenomenon. Quarterly Review of Nocturnal Imagery 3(2): 41-58.
Wren T, Hale E. (2007). Stair Count Asymmetry. Midlands Proceedings in Sleep Research 14: 112-127.
Wren T, Prasad I. (2007). Convergence Index: construction and validity. Annals of Sleep Cohort Methods 5(1): 33-49.
Kettering-Obi A. (2003). Persons in nocturnal reports: a taxonomy. Archives of Observational Dream Studies 5(1): 7-31.
LSCNI protocol, Rev. 10 (supersedes Rev. 8).

Editor's note. The authors were asked to replace the term "Chairman" with "figure 11." The authors declined, noting that the term was supplied by participants. The editor accepted the exception. Reviewer 3 comment: "I would prefer not to be named in the acknowledgments."

PAPER 4

Published 2008

Dream-Log Compliance and Waking Displacement

Okonkwo-Strand B, Pell-Kerrigan P, Hale E, Wren T. Journal of Applied Oneiric Measurement 6(3): 77-98. 2008.

In the LSCNI (n = 84 enrolled, Cohorts A-G), participants who completed 18 or more of 20 dream logs ("high compliance," n = 32) were more likely to report that objects in their sleeping rooms had been displaced on waking (17 of 32, 53%) than participants with fewer completed logs (0 of 52). Where measured (pilot photographs, n = 14 events), displacements were 2-7 cm. Ten highly compliant participants' logs also contained marginal annotations in a second hand. The association between compliance and displacement was not explained by sleep duration, age, or cohort. The findings are descriptive.

Open full paper

Archive copy · Paper 4

Received 14 April 2008. Accepted 02 July 2008.

1. Introduction

The LSCNI protocol specifies that each participant complete a dream log (Form 7-C) within five minutes of each wake prompt. In reviewing the first six cohorts, the study coordinator noted that participants who completed the log reliably tended also to mention small changes in their rooms. These were first recorded as ordinary remarks ("my shoes were by the door, not under the bed"). The present paper analyzes these remarks and the associated measurements.

2. Methods

Participants. Eighty-four adults enrolled across seven cohorts of approximately twelve (Table 1).

Table 1. Cohorts, enrollment, and consent documentation

Cohort	Dates	Enrolled	Consent forms on file
A	Oct-Dec 2004	12	12
B	Mar-May 2005	12	12
C	Oct-Dec 2005	12	12
D	Mar-May 2006	12	11
E	Oct-Dec 2006	12	11
F	Sep-Nov 2007	12	10
G	Feb-Apr 2008	12	11
Total		84	79

Compliance. The number of completed logs (of 20) was recorded for each participant. Participants were grouped as low (≤12), moderate (13-17), or high (≥18).

Displacement. After each night, participants completed a one-line item on Form 7-D: "Anything different about your room?" For a pilot subset (n = 14), participants photographed marked objects against a ruler before and after sleep (Fitch & Okonkwo-Strand, 2008).

Log examples. The logs are identified by LOG number; the first four digits correspond to sequential order across the study (LOG-0001 to LOG-1680).

3. Results

Table 2. Compliance and displacement reports

Compliance group	n	Reported displacement	%
Low (≤12 logs)	27	0	0
Moderate (13-17)	25	0	0
High (≥18)	32	17	53

Displacement was not associated with age, sleep duration, or cohort. In the 14 pilot events, displacements measured 2-7 cm (median 4); no event exceeded 7 cm. Twelve of the fourteen were toward the sleeping room's door.

Log excerpts.

LOG-0417 (S-021, night 17): "Counter 6. The Clerk handed me a photograph of the kitchen. I don't know whose kitchen. Slippers moved this morning, about a hand's width."
LOG-0422 (S-022, night 2): "Someone measured the room. The Surveyor said 'three hundred and eleven.' The bed was not where I left it."
LOG-0561 (S-029, night 1): "I was given a slip. I signed it before I was asked."
LOG-0730 (S-037, night 10): "The Usher counted me in. I counted along."
LOG-0941 (S-048, night 1): "Wrote this down. It was already written."

Annotations. In ten high-compliance participants (S-021, S-022, S-029, S-037, S-048, and five others), twenty-three log entries carried short marginal annotations ("Reviewed," "Correction: north," "Item 4 should read 3") written in a second hand: in red grease pencil in twenty entries, in blue-black ink in three. Neither was the participant's. The annotations are discussed in Hale et al. (2009).

4. Discussion

The association between completed logs and reported displacement is, as far as we can determine, not an artefact of sleep or of cohort. A selection effect (conscientious participants may notice more) is possible but would not explain the absence of any report in 52 participants.

We are not able to account for the direction of displacement, which was toward the door in 12 of 14 events. We note that displacement was never reported in a participant who had completed fewer than 18 logs, and that the threshold is sharp: participants who completed 17 logs reported none.

The annotations present a separate problem. They appeared only in the logs of highly compliant participants. They were written after the participants had left the laboratory, according to the sealing records. The logs were stored in a locked cabinet to which only the study coordinator and the principal investigator had keys.

Protocol note. This analysis follows protocol Rev. 10 (supersedes Rev. 8). Rev. 9 could not be located in the register at the time of writing.

Limitations. Displacement was reported by participants in the main sample and measured only in a pilot subset. The pilot measurements were photographic. The sample is small. Consent documentation is reported in Table 1 as it stands in the register.

Acknowledgments

Supported by the Ardent Foundation (AF-0412), the Tindall Family Trust, and a pilot grant from the Department of Behavioral Health (BH-P-2217). We thank P. Pell-Kerrigan, who hums, and who is thanked for it.

References

Hale E, Prasad I, Okonkwo-Strand B. (2006). The Waiting Room Phenomenon. Quarterly Review of Nocturnal Imagery 3(2): 41-58.
Wren T, Hale E. (2007). Stair Count Asymmetry. Midlands Proceedings in Sleep Research 14: 112-127.
Haddad-Lindqvist N, Fitch A, Hale E. (2007). Attendant Roles in Reported Nocturnal Figures. Archives of Observational Dream Studies 9(4): 203-226.
Fitch A, Okonkwo-Strand B. (2008). Household object displacement: a pilot measurement. Neurobehavioral Methods Letters 4(2): 61-68.
Hale E, Prasad I, Okonkwo-Strand B, Marsh-Teague J, Wren T. (2009). Evidence for Reciprocal Annotation. Quarterly Review of Nocturnal Imagery (withdrawn).
Lindgren-Abara S. (2005). Sleeping-room changes and report bias. Journal of Applied Oneiric Measurement 3(1): 19-34.

Reviewer comments (Reviewer 1, excerpt). "Table 1 does not sum. Eighty-four are enrolled; seventy-nine forms are on file. I raise this as a clerical point." — Author response: "The table is reproduced from the register." — Reviewer 1 (round 2): "Noted."

Withdrawn

PAPER 5

WITHDRAWN

Evidence for Reciprocal Annotation

WITHDRAWN. Quarterly Review of Nocturnal Imagery, issue 5(4), 2009.

Open full paper

Archive copy · Paper 5

5a. Withdrawal notice

The article "Evidence for Reciprocal Annotation" by Hale E, Prasad I, Okonkwo-Strand B, Marsh-Teague J and Wren T (accepted 28 August 2009) has been withdrawn before issue, at the request of the editors and with the agreement of the corresponding author (Hale). Following acceptance, the editors observed that supplementary scans submitted with the article were not identical to the scans in the manuscript as accepted. The authors state that the scans in their possession are identical to the versions they originally submitted. The editors have been unable to reconcile these statements. The corresponding author did not dispute the withdrawal. The article will not appear in a printed issue. This notice is dated 14 December 2009.

Document ID: QRNI-W-5-4-014. Revision of this notice: 2. Revision 1 is not available.

5b. Surviving author draft (file: annot_draft_v3(1).doc; last modified 03:07)

Abstract. In the LSCNI, twenty-three entries in the dream logs of ten highly compliant participants carry marginal annotations in a second hand. Annotations corrected a specific detail (a direction, a number, a counter). In 21 19 of 23 cases, the corrected detail was reproduced by the participant in a later log, without access to the annotation (logs were sealed before any participant could view them). The probability of this pattern under random agreement is ~~p < 0.001~~ p = 0.04. We call this reciprocal annotation.

1. Methods. Logs were scanned at 08:00 daily by the study coordinator and sealed. Annotations were identified by comparing the sealed originals (retrieved under supervision) with the scans. Annotation text was transcribed by two coders. Participants had no access to sealed logs.

2. Results.

Annotation type	n	Corrected detail reproduced later	Not reproduced
Direction (north/left, etc.)	8	8 7	0 1
Number (counter, step count)	9	9 8	0 1
Word or name	4	4 3	0 1
"Reviewed" only	2	n/a	n/a
Total	23	21 19	0 2

Test against chance (binomial, p₀ = 0.5): ~~p < 0.001~~ p = 0.04.

[Margin, second hand, pencil: "Corrected."]

3. Annotation material. The graphite was examined by Dr. Marsh-Teague (non-destructive, standard methods). It is consistent with a red grease pencil of the type used by the study coordinator for filing labels. (This applies to the twenty pencil entries; the three entries in blue-black ink were not analyzed.) The study coordinator stated that she did not make the annotations. A handwriting comparison was inconclusive. The annotation of the 20 October log reads: "North, not left. See floor B."

4. Discussion. If annotations were made by a member of the study team, the later reproduction by participants requires an unreported channel between team and participants. If they were not, the authors have no account. The authors considered whether the logs were altered after scanning. The scanned images and originals were compared at the time of submission and found to match. The authors note that the supplementary scans no longer match one another, and that this paper has been revised by someone.

5. Conclusion. The p-values reported above are real and are given as originally calculated.

[Margin, second hand: "Reviewed. Proceed."]

References

Okonkwo-Strand B, Pell-Kerrigan P, Hale E, Wren T. (2008). Dream-Log Compliance and Waking Displacement. Journal of Applied Oneiric Measurement 6(3): 77-98.
LSCNI protocol, Rev. 8; Rev. 9 [document not found]; Rev. 10.
Marsh-Teague J. (2009). Graphite analysis of study logs. Internal technical note.

Editor's note appended to file. "The authors' draft contains revisions in more than one typeface. The editors did not make any of them."

PAPER 6

Published 2006

Log Latency as a Predictor of Convergence

Prasad I, Pell-Kerrigan P. Neurobehavioral Methods Letters 2(4): 15-22. 2006.

Abstract. The LSCNI protocol requires a dream log within five minutes of waking. Using 960 logs from Cohorts A-D (n = 48), we examined whether the time between wake prompt and first written word (log latency) predicted the degree to which a participant's account overlapped with those of others (convergence index, per participant). Latency ranged from 14 seconds to 11 minutes (median 2 min 40 s). Shorter latency was associated with higher convergence (Spearman ρ = -0.52, n = 48). The association was strongest for the first line of the log; later lines showed no association. Participants with latency under one minute (n = 9) produced accounts that overlapped by 0.78 on average, the highest observed in the cohorts. We recommend a strict five-minute window for all further cohorts. We do not recommend a shorter window: in a subsample (n = 6) allowed to write immediately, several participants wrote entries before the wake prompt had been delivered. The timestamps were checked against the polysomnography record. They were not in error.

Citation: Prasad I, Pell-Kerrigan P. Log latency as a predictor of convergence. Neurobehavioral Methods Letters 2006; 2(4): 15-22.

PAPER 7

Published 2007

Convergence Index: Construction and Validity

Wren T, Prasad I. Annals of Sleep Cohort Methods 5(1): 33-49. 2007.

Abstract. We describe the construction of the convergence index (CI) used throughout the LSCNI. The CI is the proportion of checklist features shared between two participants' accounts, computed for each pair of participants who report a building (31-item checklist, weighted by the inverse of each feature's frequency in a control corpus). In pilot validation, the index separated true matches from random pairs with an area under the curve of 0.93. Reliability across two coders was high (ICC = 0.88). Across Cohorts A-E (n = 60, 40 reporters), the mean CI was 0.69, compared with 0.18 in a control task (n = 20) in which volunteers were asked to imagine a waiting room, and 0.21 in a second control in which they were asked to describe an office. The index is sensitive to rare features: a single shared unusual feature raises the score more than several shared common ones. Three features contributed disproportionately to the observed convergence: the number display, the "slip," and a clock showing 03:07. The index does not measure accuracy, because no external standard exists. The authors thank Mr. Fitch for the second set of codings.

Citation: Wren T, Prasad I. Convergence index: construction and validity. Annals of Sleep Cohort Methods 2007; 5(1): 33-49.

PAPER 8

Published 2007

REM Fragmentation Across the 20-Night Protocol

Prasad I, Marsh-Teague J. Midlands Proceedings in Sleep Research 15: 9-24. 2007.

Abstract. Polysomnographic records from 1,200 nights (Cohorts A-E) were examined for changes in REM architecture across the 20-night protocol. REM fragmentation (number of arousals per REM period) rose modestly across nights 1-8 and then stabilized. No difference was found between participants who reported a building and those who did not. A secondary analysis examined the timing of spontaneous terminal awakenings (awakenings not preceded by a prompt). Of 161 such events, 47 (29%) occurred between 03:05 and 03:09; approximately 4.5 were expected under a uniform distribution over the period 02:00-05:00. The clustering was present in all five cohorts. The clock in Room 3 was replaced twice during the study period; the clustering was unchanged. The authors recommend caution in interpreting the cluster as a measurement artefact, since the waking times were recorded by the polysomnograph, not by a clock. No participant was aware of the pattern at the time. This paper reports no effect on any measure of daytime function.

Citation: Prasad I, Marsh-Teague J. REM fragmentation across the 20-night protocol. Midlands Proceedings in Sleep Research 2007; 15: 9-24.

PAPER 9

Published 2008

Household Object Displacement: A Pilot Measurement

Fitch A, Okonkwo-Strand B. Neurobehavioral Methods Letters 4(2): 61-68. 2008.

Abstract. Participants in the LSCNI occasionally reported that objects in their sleeping rooms had moved by morning. To measure this, fourteen participants (high-compliance subset of Cohorts E-G) photographed three marked objects against a printed ruler each evening and each morning for the duration of the protocol (20 nights). Photographs were measured by two coders blind to night and participant. Of 840 object-nights, 14 showed displacement beyond measurement error (±0.4 cm). Displacements ranged from 2.1 to 6.9 cm (median 4.0). All fourteen events occurred in different participants. Twelve moved toward the room's door; two moved away from it. In ten of the fourteen, the displaced object was the one nearest the participant's head. The measurement method has limitations: participants positioned their own objects, and the rulers were printed on standard paper. We could not find a mechanical explanation for the pattern. We propose a larger pilot with fixed cameras (not approved at the time of writing).

Citation: Fitch A, Okonkwo-Strand B. Household object displacement: a pilot measurement. Neurobehavioral Methods Letters 2008; 4(2): 61-68.

PAPER 10

Poster 2008

Reflection Viewing in the Chairman Reports (Poster)

Hale E, Fitch A, Marsh-Teague J. Poster P-31, Fifth Annual Midlands Conference on Sleep and Perception, 2008.

Poster text.

Background. Of 84 LSCNI participants, 19 (23%) described a figure with a headwear stimulus (HWS). Method. All 19 accounts were examined for the mode of viewing. Results. In 19 of 19 accounts, the figure was seen only in a reflective surface. In none was the figure seen directly. Surfaces included: polished counter (7), window glass (5), framed notice (4), a participant's own watch face (2), and an unspecified "shine" (1). In four participants with multiple reports (S-017, S-031, S-044, S-062), no description of the figure was repeated across nights. Participants' own wording: "He is whoever is in the glass." "When I turned, there was a coat on a hook." "The Ushers did not look at the glass."

Cohort G (n = 12): mirrors were placed in two of four laboratory rooms (Rooms 2 and 4). Of the five HWS reports in Cohort G, three came from mirrored rooms (not significant, Fisher p = 0.99). Interpretation. None offered. Note to viewers: please do not photograph the poster. Poster board coordinates: the poster was displayed for two days. On the second day, the board was found to be turned to face the wall.

Citation: Hale E, Fitch A, Marsh-Teague J. Reflection viewing in the Chairman reports. Poster P-31, Fifth Annual Midlands Conference on Sleep and Perception; 2008.

Retracted

PAPER 11

RETRACTED

Affective Correlates of Recurrent Setting Reports

Haddad-Lindqvist N. Journal of Applied Oneiric Measurement 7(1): 12-30. 2009.

Abstract (original). Participants who reported the same building on ten or more of twenty nights (n = 29, Cohorts A-G) rated their familiarity with their immediate surroundings on waking more highly on a 7-point non-clinical scale (mean 5.8) than participants who reported the building on fewer nights (n = 55, mean 4.1; difference 1.7, 95% CI 1.2-2.2). The difference was largest in the first five minutes after waking and diminished over the following hour. No clinical measure was taken. The author makes no claim about wellbeing or health. The scale items concerned recognition of the room, of furniture, and of the location of doors. Rater agreement was high. The study has a single author and a single site.

RETRACTION NOTICE

Retraction: "Affective Correlates of Recurrent Setting Reports." Journal of Applied Oneiric Measurement 7(1). Notice dated 03 June 2009. Notice no. RN-2009-006 (RN-2009-005 not issued).

This article has been retracted at the request of the Human Subjects Review Committee of Halloran-Reeve University (Chair: Prof. C. Ridley-Bane) and the editor. The retraction follows an audit of the consent documentation for Cohorts D-G, in which the number of consent forms on file did not match the number of participants. The audit further found that the scale data for 16 participants could not be matched to a source form. The article therefore does not meet the journal's requirements for documented consent and verifiable data.

The author was informed of the retraction on 02 June 2009 and provided the following statement, which the editor has agreed to publish:

"I accept that the documentation is incomplete. I do not accept that the data are wrong. The author does not withdraw the effect."

The retracted article remains online, watermarked on each page as "Retracted," to maintain the scholarly record. Document ID: JAOM-R-7-1-006 (notice no. RN-2009-006).

Citation: Haddad-Lindqvist N. Affective correlates of recurrent setting reports. Journal of Applied Oneiric Measurement 2009; 7(1): 12-30. Retracted 03 Jun 2009.

Not issued

PAPER 12

NOT ISSUED

Log Closure and Subsequent Reports

Hale E, Prasad I, Okonkwo-Strand B, Wren T. Journal of Applied Oneiric Measurement (in press). Manuscript ID: JAOM-2009-0148.

Abstract (as submitted). In the final week of the LSCNI protocol, thirty-one participants were asked to close their dream logs with a single line, "Done," on the final night. Subsequent contact (a follow-up questionnaire at six weeks, n = 31) asked whether the participant had continued to see the setting. Twenty-eight participants reported no further setting reports. Three reported that the setting continued. In all three, the final line of their log had been annotated in a second hand with an additional line. The annotations read: "Not yet." (S-029), "Noted." (S-048), and "Proceed to Floor B." (S-021). These three participants' logs were the only ones in the archive in which the closing line had been altered. A fourth annotation, on the log of a participant who did not respond to follow-up (S-037), has been withheld from this abstract.

We recommend that the protocol's closing procedure not be applied in any further study.

Status: not issued.

Editorial history: Accepted 21 October 2009. Typeset 03 November 2009. Proof sent to authors 17 November 2009. Proof returned 18 November 2009 by the corresponding author with the marginal note "Please do not run this." The editor wrote on 27 November that the article was withdrawn from production at the editor's discretion, with the reason recorded as "see correspondence." The correspondence has not been located.

Staff

Halloran-Reeve University, Department of Neurobehavioral Medicine. Records extract, JAN 2010. Eight staff files. Photographs not retained.

HR-0314-H

RESIGNED

Hale, Elias, Ph.D.

Principal Investigator

Personnel status: RESIGNED, effective 30 NOV 2009. Exit interview not held.
Employee no.: HR-0314-H. Years of service: 11.

Open personnel file

Dr. Hale directed MSSPL from its opening and designed the Longitudinal Study of Convergent Nocturnal Imagery (LSCNI). His earlier work concerned reaction-time variance in shift workers. Colleagues describe him as methodical and unhurried. He stirred his coffee seven times, always seven, and then set the spoon on the saucer with the bowl pointing away from him. Once, when a coder asked why, he said it kept the count.

He is the author or co-author of seven of the twelve listed publications. His signature appears on Consent Form v.4 and on the withdrawn paper. In October 2006 he left a blank log overnight in Room 3 (LOG-1681, numbered after the series). In the morning, in his handwriting, it read "Reviewed. Proceed to Floor B." He did not discuss the line again, but is recorded as remarking that it had been "a good hand."

Excerpt: resignation letter, 02 NOV 2009

To the Dean, Faculty of Medicine and Health Sciences.

I am writing to resign my appointment as Principal Investigator, Marsh Street Sleep and Perception Laboratory, and as Reader in Neurobehavioral Medicine, effective 30 November. I have enjoyed the support of the Department for eleven years.

I do not offer a reason beyond the obvious one, which is that the Laboratory has been archived and my continued presence is no longer required. Ms. Okonkwo-Strand has the log index. Dr. Haddad-Lindqvist has the retraction file. I have left the sleeping rooms as I found them, except for Room 3, which I have not been able to enter since 14 October.

I would ask that the keys not be re-cut. I understand they are going to be. I ask that Room 3 not be booked past the 14th.

Yours faithfully,

E. Hale

Quirk and fate: Hale never signed the departure register. His unsent last email is held as Document D-5. Forwarding address: none supplied.

HR-0587-P

TRANSFERRED

Prasad, Imogen, M.D., Ph.D.

Co-Principal Investigator, Polysomnography

Personnel status: TRANSFERRED to "another institution," 15 MAR 2009. Forwarding address: none.
Employee no.: HR-0587-P.

Open personnel file

Dr. Prasad built the overnight recording protocol: scalp EEG, eye movement, and chin EMG on all four beds, with wake prompts at REM onset. She was particular about electrode paste and about the cap numbering, which she insisted be written inside the cap, not on the outside. Her recordings are the cleanest in the archive. They also carry, in every record from 2005 on, a flat segment on all channels from 03:07:00 to 03:07:59 that she attributed to the amplifiers. Three amplifiers were replaced. The segment remained.

She is recorded as the first to ask, in writing, for the corridor camera. She is also recorded as the first to ask that it be turned off.

Excerpt: laboratory notebook, 12 NOV 2005

Room 4, night 12. Subject asleep 23:48. REM onset 01:21. Wake prompt delivered, subject wrote 41 words.

Study clock and bedside clock disagree by 4 min 12 s. Offset constant for the third night. Checked at start and end: same.

03:07: all channels flat, 59 s. Amplifier 2 swapped. Flat again, same minute, on the new unit.

Corridor camera shows the Room 3 door ajar at 03:07. Door log shows no entry. I have not mentioned this to Elias. He would ask me to count something.

Subject, in the morning: "I was told you'd been very thorough." Said it politely.

Quirk and fate: Dr. Prasad counted the electrodes on every cap twice and the chairs in the waiting area outside Room 1 three times. She left the university on 15 MAR 2009. Her office was found cleared, with a note taped to the drawer: "Not hers. Not returned." Payroll lists her as transferred. The receiving institution is unnamed.

HR-0611-W

LEFT THE FIELD

Wren, Tobias

Statistical Analyst

Personnel status: LEFT THE FIELD. Contract not renewed; resignation accepted 30 JUN 2008.
Employee no.: HR-0611-W.

Open personnel file

Mr. Wren ran the convergence index and the log-latency models. In 2007 he noticed that in all 31 reports giving both counts, stairs counted in descent exceeded those counted in ascent by exactly one. The variance of the difference was zero. A fit was not possible. He re-ran the analysis five times.

His colleagues recall that he stopped using the stairs in the Faculty building and took the lift for the last year of his contract. He told a coder that he no longer cared to "have the number confirmed." He is now believed to work outside statistics. No one has located his address.

Excerpt: email to E. Hale, 11 MAR 2007, 22:14

Elias,

I have re-run the stair counts. Descent minus ascent equals 1 for 31 of 31 reports. I checked the scans for transcription slips and found none.

Spread is zero. This is not noise and it is not bias in the usual sense; there is nothing to be biased. If a rater were guessing we would see spread. There is none.

I do not think I can write this up in the usual way. I do not think I can write it up in any way that survives review, which, as you know, includes Professor Ridley-Bane.

I'll draft it. I'd ask that my name come second.

Tobias

P.S. I counted the stairs to the third floor this evening. Going up, 31. Going down, 31. I recounted. 31 and 31. I do not know what I expected.

Quirk and fate: Wren hummed in the rhythm of his counting. "Stair Count Asymmetry" (2007) carries his name first, against his request. After 2008 his profile on the Department page reads: "Tobias Wren has left statistics." The sentence is not a reference to any appointment.

HR-0702-HL

PAPER RETRACTED

Haddad-Lindqvist, Noor, M.D.

Psychiatrist, Clinical Liaison

Personnel status: PAPER RETRACTED 03 JUN 2009. Contract ended 31 AUG 2009.
Employee no.: HR-0702-HL.

Open personnel file

Dr. Haddad-Lindqvist saw participants before and after each cohort and added notes to selected logs alongside the formal coders (Fitch and Prasad). She asked each participant the same closing question: "Did anyone ask you to sign anything?" She recorded the answers on a card she kept in her breast pocket and did not enter into the central file.

In May 2008 she submitted "Affective Correlates of Recurrent Setting Reports," which reported that participants who described the same building on ten or more nights rated their familiarity with their surroundings on waking more highly. In June 2009 she retracted it. Her notice (D-2) states that she does not withdraw the effect.

Excerpt: interview note, 04 JUN 2008

S-079, follow-up interview, 40 min. Calm. No distress.

S-079: "They're very kind about it. They correct the log and they're always right. That's what I find difficult. I haven't had a thing to disagree with."

I asked whether the corrections were frightening. S-079 said no, they were polite.

I asked whether she would continue keeping logs. S-079 said she had been asked to. She used the word "asked."

Me: Did anyone ask you to sign?

S-079: "I think I'd already signed."

Card updated. No clinical concern. Personal concern: noted, not entered.

Quirk and fate: Dr. Haddad-Lindqvist carried her notes in a green folder and never in a bag. She closed her practice in December 2009. A letter to the Dean asks that her notes "be kept in a dry place, and not read aloud."

HR-0233-OS

RETIRED

Okonkwo-Strand, Beatrix

Study Coordinator and Log Keeper

Personnel status: RETIRED 15 JAN 2010. Floor B filing allowance continues (see note).
Employee no.: HR-0233-OS.

Open personnel file

Ms. Okonkwo-Strand received, indexed and stored every physical log. She marked them in red pencil only, never ink, and sharpened it to a stated length of four centimetres. She was the first to notice that two participants who had never met had described the same counter, the first convergence, in October 2004.

She resisted every change to the log format. She argued that logs should be read "in the order received, not the order written," and that "nobody should answer them." Her handover memo ends: "They answer. Please do not."

Excerpt: log-keeping note, 22 OCT 2004

Index card. S-006 / S-0 (Subject 0, IP-04).

Both describe a counter with a brass bell. Both: board reading 40. Neither participant knows the other. I checked the consent files, the sign-in book, and the bus pass list.

I have marked both in red. I do not like this. It is very neat.

Next to the card I have put a second card. It was not there when I went to lunch. It reads: "Noted. Thank you, B."

Quirk and fate: The Access Control office reports that her badge was deactivated in November 2009 when the keys were re-cut. It records an entry under her number every night at 03:07 at the Marsh Street building's lower door. Facilities was asked to investigate in December. No reply is on file. The filing allowance is paid to an account that does not exist; HR reports the payments are "accepted." In informal staff chat, she is described as one who "still files at the Annex."

HR-0455-MT

NO RECORD FOUND

Marsh-Teague, Julian, Ph.D.

Neuroimaging

Personnel status: NO RECORD FOUND (see stub).
Employee no.: HR-0455-MT (as quoted in correspondence).

Open personnel file

Dr. Marsh-Teague ran the imaging component of LSCNI in a small number of participants. He is remembered for two rules: scans were to be taken in the horizontal plane only, and no image was to show a ceiling. He kept his lanyard turned so the badge faced his chest. He never photographed a ceiling, including in his own flat, and he declined to go into the Laboratory's Room 3 during the day.

His name appears on three publications. In the University's current staff directory he does not. The Records Office has no entry.

Excerpt: scan protocol note, 03 FEB 2007

Protocol 2B. Acquisition angle fixed; field of view excludes superior margin. No scout images in the sagittal that include the ceiling of Suite 1. If a scout is taken, discard.

Technician queried the instruction. Reply: "It is a quality matter." Technician accepted that.

I have noticed that at the upper edge of each slice there is an area of a different grey, regular, narrow. Not artefact. I have asked the vendor. The vendor says the machine is working normally.

Records Office stub:

RECORDS OFFICE, HALLORAN-REEVE UNIVERSITY

Search: MARSH-TEAGUE, J. Date of search: 12 JAN 2010.

Result: NO RECORD FOUND.

Cross-searched: Payroll, Appointments, Parking, Library, Staff Directory.

Note: Form completed in error by requester. Requester also not found.

Quirk and fate: Removed from records, effective date unrecorded.

HR-0890-PK

TRANSFERRED

Pell-Kerrigan, Pilar

Junior Technician, Night Shift

Personnel status: TRANSFERRED to Accessions, 01 NOV 2007. Employed.
Employee no.: HR-0890-PK.

Open personnel file

Ms. Pell-Kerrigan, then 24, monitored the overnight recordings and signed the door log. Her shift log is neat, daily, and slightly amusing. On the night of 18 SEP 2007 it stops. It resumes on 02 OCT 2007, in what Ms. Okonkwo-Strand certified as the same hand.

The log contains no mention of the fortnight. The resumed entries refer to a "Floor B schedule" and to staff who are not in the personnel register. She no longer works nights.

Excerpt: night technician's shift log, 18 SEP 2007

22:00 Shift start. Rooms 1, 2, 4 occupied. Room 3 empty (booked).

23:15 Participant S-066, Room 2, asleep.

01:40 Wake prompts delivered.

02:51 Corridor quiet. Heating ticking.

03:05 Heard something in the corridor, paper shuffling.

03:07 I can hear the hat.

Resumed entry, 02 OCT 2007:

22:00 Shift start. Floor B schedule received from the Chairman's office. All quiet. No incidents. Thank you for asking.

Quirk and fate: She never wrote the word again. She now works in Accessions, Level B, with her back to the window, and asks not to be photographed.

GS-0912-F

THESIS

Fitch, Arlen

Graduate Student, Dream Coder

Personnel status: THESIS "completed elsewhere." Enrolment closed 14 DEC 2009.
Student no.: GS-0912-F.

Open personnel file

Mr. Fitch coded the logs, wrote the first definition of "convergence," and tabulated the building features. He is the author of the coding manual (Rev. 10, which supersedes Rev. 8). He asked in writing that Rev. 9 be "corrected." No Rev. 9 exists.

Mr. Fitch kept a small top hat sticker on the lid of his laptop. He said repeatedly that it was not his and had been there when he bought it. Colleagues noted that it moved between weeks, a few centimetres each time, toward the hinge.

Excerpt: coding journal, 07 OCT 2007

Recoded LOG-1305 (S-066). Chairs, nine. Reflective surface. Fire-notice frame.

I need to record, for the file, that the sticker on my laptop is not mine. It was a third-hand machine. I did not apply it. It appears to be an ordinary adhesive vinyl. It has moved again, about 3 cm.

I am not going to use the other word in the notes. I have used the approved term. I would like it noted that the term is approved.

Someone has red-penciled the corner of the page: "Noted, A." I did not put that there.

Quirk and fate: Fitch's thesis was accepted, according to the Graduate School, "completed elsewhere." The examiner is unnamed. The deposit copy is not held at any library. In a final email, he asks that the sticker be returned to its owner.

Timeline

Dates as they appear in the documents, logs and personnel files.

Feb 2004 Insomnia Pilot IP-04 enrolls Subject 0, a night nurse, on 03 February. Her third log (17 February) describes an annex to the ward that is not on the plan. File #oyelaran.
Oct 2004 First convergence noticed. On 22 October S-006 describes the same counter and the same board as Subject 0, without having met her. The log keeper records that two participants who had never met have described the same room.
Feb 2007 MEMO-0417-K (14 February): the word "hat" is not to be used on coding sheets. The abbreviation HS is introduced.
Apr 2008 The final participant, S-084, completes night 20 on 22 April. Enrolment is closed.
Oct 2008 Consent concerns. The Human Subjects Review Committee (22 October) counts 79 signed forms against 84 enrolled participants.
Feb 2009 Tindall pause. The Tindall Family Trust (17 February) pauses grant payments.
Mar 2009 A press release is drafted on 9 March and not issued. Dr. Prasad is transferred on 15 March.
Jun 2009 Paper 11 is retracted (notice RN-2009-006, 03 June).
Aug 2009 Paper 5 is accepted on 28 August and withdrawn before issue. Dr. Haddad-Lindqvist's contract ends on 31 August.
Oct 2009 Dr. Hale is unable to enter Room 3 after 14 October. Paper 12 is accepted on 21 October.
Nov 2009 Keys re-cut. Dr. Hale resigns (letter dated 02 November, effective 30 November). A draft email is saved unsent on 13 November. The stock-take of 20 November finds one more chair than the register. Inventory INV-2009-11.
Jan 2010 The laboratory is archived.

Documents

MSSPL archive, JAN 2010. Transcribed as found. Original typographical errors retained.

D-1

Internal Memorandum MEMO-0417-K

Date: 14 FEB 2007 (typed "2006", corrected by hand)
From: E. Hale, Principal Investigator
To: All coders and rating staff
Cc: I. Prasad; B. Okonkwo-Strand
Re: Hat references in coding
Page 1 of 2

Effective Monday 19 February, the word "hat" is not to be used on coding sheets, convergence forms, or in the margins of participant logs. Where a participant uses it, the coder will transcribe the entry verbatim in the log and enter the stimulus on the coding sheet as "headwear stimulus" (abbreviation HS).

This is a consistency measure. It is not a request for discussion.

Rationale (for the file only): the term "hat" has produced inter-rater disagreement in the 0.40–0.60 range, in part because it carries connotations (formal, ceremonial, theatrical) that are not present in participant reports. "Headwear stimulus" is neutral. It is also, we note, harder to say.

Coders are reminded that:

The abbreviation HS applies only to the object, not to the figure. The figure is coded under "Attendant" or "Chairman" as applicable.
Coders will not describe the figure's face. Reports that describe it are to be entered as "not stated."
Coders will not view the figure directly on any recording, photograph or reflective surface. If an image is necessary, a second rater will describe it.
Rev. 10 of the Coding Manual supersedes Rev. 8. Rev. 9 is not to be requested.

Existing sheets will be amended by B. Okonkwo-Strand in red pencil. Please do not amend your own.

Any query should be put in writing and not discussed in the corridor.

D-2

Notice of Retraction RN-2009-006

Date: 03 JUN 2009
Publication: Journal of Applied Oneiric Measurement, Vol. 7, No. 1
Retracted article: Haddad-Lindqvist, N. (2009). "Affective Correlates of Recurrent Setting Reports." pp. 12–30.

I, the undersigned, retract the above article in full.

The article reported that 29 participants who described the same building on ten or more of twenty nights rated their familiarity with their surroundings on waking more highly than the remaining 55 (5.8 against 4.1). The figures reported in Table 3 are accurate. The data have been re-examined and the numbers have not changed.

The retraction is made on the following grounds.

The title of the article uses the word "affective," which implies a state of feeling or health. The study did not examine a state and the author withdraws any inference to that effect.
The interpretation in the Discussion, that recurrent reports "settle" the waking surroundings, is withdrawn as unsupported.
The author is no longer certain who is becoming familiar with whom.

The author does not withdraw the effect.

Participants named in the article have been informed in writing. Several declined the letter on the grounds that they had received it.

D-3

Letter, The Tindall Family Trust

Ref: TFT/MSSPL/09-02
Date: 17 FEB 2009
To: Dr. E. Hale, Marsh Street Sleep and Perception Laboratory (MPSSL in header)
From: The Tindall Family Trust, 4 Alder Court

Dear Dr. Hale,

Further to the Trustees' meeting of 9 February, I am writing with regret to confirm the following.

The Trustees have reviewed the Laboratory's interim report, the supplementary correspondence from the Human Subjects Review Committee, and your letter of 12 January. We have concluded that the scope of the project no longer aligns with the Trust's priorities as set out in our 2004 grant agreement.

Accordingly, with effect from receipt of this letter, payments under the grant are paused. The second instalment of three, due on 1 March, will not be released. (We note that the instalment is, in fact, the fourth. The error is ours. The decision stands.)

Payments will be reviewed once the Trustees have received a satisfactory clarification of the matter referred to in your report as "the Annex." We would be grateful if you would confine that clarification to the plain facts, avoiding adjectives. We have asked a colleague to read the supplementary correspondence and have been told, without further detail, that it was "thorough."

The Trustees remain grateful for your work, and nothing here reflects on the integrity of the researchers. We would ask that the Trust's name not be associated with the Annex in any publication, public or private, past or planned.

With best wishes,

D-4

Memorandum, Human Subjects Review Committee

Ref: HSRC-0842
Date: 22 OCT 2008
From: Prof. Constance Ridley-Bane, Chair, HSRC
To: Dr. E. Hale; Dean's Office (Medicine)
Re: Consent concerns, LSCNI (Protocol HSRC-04-118)
Page 3 of 11 (pp. 4–9 withheld)

I have now completed my review of the consent documentation for the above study, and write to put my concerns in plain terms.

Consent Form v.4, the current version, is a four-page document describing overnight recording, wake prompts and the keeping of morning logs. It does not describe the Annex.

I do not say this as a procedural criticism. I say it because I have read the logs, and I cannot find a sentence that would describe the Annex in a form a participant could be asked to understand. The Annex is not a risk we can name.

I have counted the signed forms twice. There are 79. There are 84 enrolled participants. I have asked the Principal Investigator where the other five are. He replied, accurately, that the logs exist. I have not been told why the forms do not.

The effect, if the term is acceptable, appears only in those who complete the logs. I therefore cannot recommend any change that would encourage completion. I also cannot recommend any change that would discourage it.

I request that no new participants be enrolled pending a full review. I am aware that enrolment closed on 22 APR.

I note, for the Committee's records, that I have been asked on three occasions this month to sign things I do not recall seeing. I have declined. I mention it for completeness only.

D-5

Draft email (unsent)

From: e.hale@mssp.hr-u.example
To: [blank]
Subject: (no subject)
Saved as draft: 13 NOV 2009, 03:07
Status: Not sent

Beatrix,

(Delete this if you can't use it.)

The keys are being re-cut on the 20th. Facilities has asked for a list of rooms. Please give them one.

Room 3 should not be booked past the 14th.

I realise that sentence is unhelpful. It is the sentence I have. Please note it in the booking sheet, in your usual pencil, and ensure the next person to hold the schedule reads it as written.

The stair count is the same each time. If someone asks, give them the number and no more. Do not say we have counted it more than once.

Tell Imogen I understood her note, if you find her. Tell Tobias I'm sorry about the order of the names.

I have not signed anything. I have been told that is not how it works.

I will take the log home. I will not complete it.

Thank you for the pencil.

D-6

Press Release PR-2009-004 (not issued)

Date: 09 MAR 2009 (draft 3)
Status: HOLD. Not for distribution.

FOR IMMEDIATE RELEASE

HALLORAN-REEVE UNIVERSITY ANNOUNCES RESULTS OF LONGITUDINAL SLEEP STUDY

Marsh Street, 9 March. The Department of Neurobehavioral Medicine at Halloran-Reeve University today announced the completion of data collection in the Longitudinal Study of Convergent Nocturnal Imagery (LSCNI), a five-year investigation of █████████ in 84 volunteers.

"We are delighted to report that the study found ████████ in ██ of 84 participants," said Dr. Elias Hale, Principal Investigator. "Most striking was the consistency of the ███████████, described independently by participants who had never met. We believe this has implications for ██████████████."

Participants kept nightly logs for 20 nights. Of these, ██ reported a ███████ ██ █████, and 61 described the same ████████. Researchers also recorded measurable ██████████ of household objects, ranging from 2 to 7 centimetres, in ██████ ██ ███████ ████.

"The Annex is not ████ ███ ███████," said Dr. Hale. "It is ██████ ██████."

The study was supported by the Tindall Family Trust, the Ardent Foundation for Cognitive Research, and a behavioral-health pilot grant. ██████████ ████████ ██ ██████.

Participant logs

Marsh Street Sleep and Perception Laboratory. Archived JAN 2010. 33 of 1,680 numbered logs shown (1,642 originals located; the 38 not located survive as scans only). Log numbers run in participant order: LOG-0001 is S-001, night 1; LOG-1680 is S-084, night 20 (LOG = 20 x (S - 1) + night). Shown here in order of writing. Participant codes S-001 to S-084. Consent version (Form 4-C, v.1 to v.4) is shown where one exists.

Coder notes dated before MEMO-0417-K (14 FEB 2007) refer to the figure in the participant's own words. Coder notes after use "headwear stimulus" (HS).

Cohort A · S-001 to S-012 · 21 OCT 2004 to 30 NOV 2004 · 3 logs shown

LOG-0102 · S-006 · 21 OCT 2004 · Night 2 of 20 · Consent v.1

Retired electrician, 63. "Waiting somewhere with a lot of chairs. Not a station. Brown carpet with a pattern, like the old council offices. A counter with a brass bell and a flip board above it showing 40. Nobody serving. I sat down and felt I had been there since before I fell asleep. Woke at 3:07 by the bedside clock. I checked it twice."

Coder note (A.F., 02 DEC 2004): Waiting area, counter, board "40." Unprompted.

Convergence flag: B1 (building, partial). First "40."

LOG-0165 · S-009 · 02 NOV 2004 · Night 5 of 20 · Consent v.1

Graduate student, 24. "Same waiting room as the last two nights, though I would not have said so before tonight. Coat rack by the door. Someone had hung my coat on it, the green one, with the collar turned out the way I never wear it. I did not take it down. It did not seem to be mine to take."

Coder note (A.F., 02 DEC 2004): Coat on hook. Participant states coat is hers.

Convergence flag: B1.

LOG-0209 · S-011 · 30 NOV 2004 · Night 9 of 20 · Consent v.1

Bus driver, 51. "Counted chairs because there was nothing else to do. Nine. Nine every time. A man at the counter asked whether I was here for the first appointment or the second. I said I didn't know. He said that was fine, that most don't, and put a mark in a book. He did not hurry. Nobody in there hurries."

Coder note (B.O-S., 05 JAN 2005): Nine chairs again (cf. S-006, S-009 estimates). Attendant, counter, ledger.

Convergence flag: B2.

Cohort B · S-013 to S-024 · 24 MAR 2005 to 05 MAY 2005 · 4 logs shown

LOG-0264 · S-014 · 24 MAR 2005 · Night 4 of 20 · Consent v.1

Accountant, 38. "Counter again. The clerk turned the book round so I could see my name in it. Spelled right, which nobody manages. Beside it a column of dates, some in the future. I asked about those. He said they were provisional. I woke with the pen from the bedside table in my hand, though I'd put it away."

Coder note (A.F., 22 APR 2005): Ledger with participant's name; future dates. Pen displacement reported, unverified.

Convergence flag: B2.

LOG-0422 · S-022 · 29 MAR 2005 · Night 2 of 20 · Consent v.1

Pastry chef, 29. "Someone measured the room. The Surveyor said 'three hundred and eleven.' The bed was not where I left it."

Coder note (A.F., 25 APR 2005): Surveyor (participants' term, provisional). Figure spoken aloud, centimetres assumed. Bed displacement not measured.

Convergence flag: S2 · D1.

LOG-0431 · S-022 · 22 APR 2005 · Night 11 of 20 · Consent v.1

Pastry chef, 29. "A staircase behind the counter, narrow, carpeted, going down. Counted sixteen steps going down. I came back up, I don't know why, to check, and it was fifteen. I did it again. Same. Nobody stopped me, and a man at the top watched without any interest at all. Going down is longer. I can't make it sound reasonable."

Coder note (A.F., 06 MAY 2005): Stair count. Down 16, up 15. Compare S-017's count (Cohort B, earlier night).

Convergence flag: B2 · C1 (count).

LOG-0417 · S-021 · 05 MAY 2005 · Night 17 of 20 · Consent v.1

Claims adjuster, 37. "Counter 6. The Clerk handed me a photograph of the kitchen. I don't know whose kitchen. Slippers moved this morning, about a hand's width."

Coder note (B.O-S., 19 MAY 2005): Clerk shows an item; kitchen not recognized. Slipper displacement is the participant's estimate, unmeasured.

Convergence flag: B2 · D1.

Cohort C · S-025 to S-036 · 18 OCT 2005 to 02 DEC 2005 · 6 logs shown

LOG-0561 · S-029 · 18 OCT 2005 · Night 1 of 20 · Consent v.2

Dental receptionist, 31. "I was given a slip. I signed it before I was asked."

Coder note (N.H-L., 14 NOV 2005): Signature precedes the request. Night 1.

Convergence flag: B1 · S1 (signature).

LOG-0602 · S-031 · 25 OCT 2005 · Night 2 of 20 · Consent v.2

Taxi dispatcher, 44. "The board over the counter said 40. It always says 40. I don't know how I know that, it's only my second night. A woman beside me held a ticket with 41 on it. She said she'd been waiting a long time. The board didn't change while I watched. I've been awake an hour and I keep looking at the clock to see if it says 3:07. It does."

Coder note (A.F., 21 NOV 2005): Board "40"; ticket "41." Participant reports having "known" the number. No prior log contact.

Convergence flag: B2 · N1 (number).

LOG-0486 · S-025 · 03 NOV 2005 · Night 6 of 20 · Consent v.2

Retired schoolteacher, 67. "Window at the far end of the waiting room, tall, with a blind. In the glass, not through it, there was someone standing behind my chair. Taller than a person should stand in a room that low. I could only see him in the reflection. I turned round and the chair was empty. In the window he was still there, facing the other wall."

Coder note (B.O-S., 29 NOV 2005): Figure in window reflection only. Facing away.

Convergence flag: H1 (figure with headwear, partial; headwear not stated).

LOG-0648 · S-033 · 10 NOV 2005 · Night 8 of 20 · Consent v.2

Care assistant, 48. "Waiting room. Number 40 on the wall. Usher asked if I'd waited long. I said yes. He wrote that down. He said the next stage would be quick and that I'd be shown something of mine. I said I'd rather not. He wrote that down too. I woke before they showed me. The ceiling of my room had the same stain pattern as theirs."

Coder note (A.F., 12 DEC 2005): Usher; number 40; refusal recorded. Matches LOG-0102 (S-006) in counter and board, no contact between participants.

Convergence flag: B3 (strong).

LOG-0708 · S-036 · 10 NOV 2005 · Night 8 of 20 · Consent v.2

Plumber, 55. "Same counter. A clerk showed me my own coat on a hook. I told him it wasn't mine. He said, not yet, and made a note. Then he showed me a drawer of keys, mine among them, though I only own two. The man behind me in the queue never turned round. I could see him in the brass of the bell. He seemed to be looking at the door."

Coder note (A.F., 12 DEC 2005): Same night as LOG-0648, different room. Coat on hook (cf. S-009). Figure in reflective brass.

Convergence flag: B3 · H1.

LOG-0554 · S-028 · 02 DEC 2005 · Night 14 of 20 · Consent v.2

Library assistant, 26. "Someone with a tape measure. He measured the wall by my chair and then the chair, and wrote on a card. The tape had no numbers on it, I looked. He measured my arm last, politely, as if asking permission. In the morning the bookcase in my flat stood a few centimetres off the wall. I measured. Four. There's a dust line."

Coder note (A.F., 19 DEC 2005): Surveyor (provisional label). Displacement 4 cm, participant-measured.

Convergence flag: B2 · D1 (displacement).

Cohort D · S-037 to S-048 · 14 MAR 2006 to 06 MAY 2006 · 6 logs shown

LOG-0941 · S-048 · 14 MAR 2006 · Night 1 of 20 · Consent v.3

Hospital porter, 44. "Wrote this down. It was already written."

Coder note (A.F., 11 APR 2006): Night 1. Entry complete before the wake prompt was delivered (cf. Paper 6, log latency).

Convergence flag: B1.

LOG-0863 · S-044 · 21 MAR 2006 · Night 3 of 20 · Consent v.3

Warehouse night supervisor, 47. "Two of them at the door of the waiting room, standing either side, nothing in their hands. They didn't touch me. They didn't need to. Nobody ran when I tried the handle. They walked, quite fast, to be where I was. I stopped. They stopped. One said I'd be called shortly and the other said the first one was always optimistic."

Coder note (A.F., 14 APR 2006): Two Attendants at exit. Pursuit at walking pace. "Nobody ran" matches S-011, S-025.

Convergence flag: B2 · W1 (door-keepers).

LOG-0730 · S-037 · 05 APR 2006 · Night 10 of 20 · Consent v.3

Railway signaller, 57. "The Usher counted me in. I counted along."

Coder note (A.F., 02 MAY 2006): Usher counts participant in; the number is the participant's own.

Convergence flag: B2.

LOG-0847 · S-043 · 15 APR 2006 · Night 7 of 20 · Consent: none on file

Mother of two, 33. "The waiting room. A man in a dark coat gave me a form on a clipboard and held out a pen. I said I had nothing to sign, I'd signed already, at the university. He said, not that one, and smiled as if I had been very helpful. He said it was already done and the pen was only for the look of it. I put my name down anyway. It was already written."

Coder note (N.H-L., 02 MAY 2006): Signature request in the dream. See administrative note on consent: none supplied.

Convergence flag: B2 · S1 (signature).

LOG-0752 · S-038 · 17 APR 2006 · Night 12 of 20 · Consent v.3

Carpenter, 59. "Stairs again. Seventeen down, sixteen up. I trust my counting, I count for a living. I wrote it on the back of my hand in the dream so I wouldn't forget, and the back of my hand was bare this morning."

[Second hand, blue-black ink, in margin: "Corrected: 17 down, 16 up. Counted as stated."]

Coder note (B.O-S., 02 MAY 2006): Margin note not in participant's or coder's hand. Source of annotation not established.

Convergence flag: C2 · A1 (annotation).

LOG-0815 · S-041 · 06 MAY 2006 · Night 15 of 20 · Consent v.3

Nurse's aide, 36. "Counted the stairs down. Eighteen. Up, seventeen. Counted again. Same. At the bottom there was a door with a letter on it. B. I did not go through. A man at the top said that was quite all right, that nobody does the first time, and that the door would still be there."

Coder note (A.F., 29 MAY 2006): Stair count, 18/17. First "B" door. Independent of LOG-0431 (S-022).

Convergence flag: C3 · F1 (Floor B, first mention).

Cohort E · S-049 to S-060 · 30 OCT 2006 to 18 DEC 2006 · 6 logs shown

LOG-1681 · STAFF (Hale, E.) · 30 OCT 2006 · Room 3, unobserved · Consent: n/a

[Log form left on bedside table at 23:41. Entry as found at 06:52: blank.]

[Beneath, in the principal investigator's own handwriting, in ink not issued by the laboratory:] "Reviewed. Proceed to Floor B."

Coder note (B.O-S., 31 OCT 2006): Form issued blank. Handwriting examined against PI's sample: consistent. PI states he did not write the line. Accessioned late (NOV 2006) at PI's request, and numbered after the series (LOG-1680 was already assigned).

Convergence flag: none assigned. Room 3 had been booked past the 14th.

LOG-1086 · S-055 · 02 NOV 2006 · Night 6 of 20 · Consent v.3

Retired postmaster, 70. "Surveyor measured my bed. He said, 'You'll be sleeping here later.' It's my own bed, at home, so I took that to be a mistake. He measured the door as well, and the gap beneath it. In the morning the wardrobe had moved. Not much. Enough that the door catches."

Coder note (A.F., 30 NOV 2006): Surveyor measures a room to be occupied. Wardrobe displacement, est. 6 cm.

Convergence flag: S2 · D2.

LOG-1049 · S-053 · 14 NOV 2006 · Night 9 of 20 · Consent: none on file

Security guard, 52. "I watch screens all night, that's my job, so I notice reflections. In the lift doors there was a shape standing just behind my left shoulder, a darker place in the steel, quite tall, quite still, and it never turned its head toward me. I looked straight at the man beside me. Nothing. I looked at the doors. There. I did not look again."

Coder note (N.H-L., 02 DEC 2006): Reflection only. No consent form retrievable; participant completed Nights 1-20.

Convergence flag: H2 · R1 (reflection).

LOG-1030 · S-052 · 22 NOV 2006 · Night 10 of 20 · Consent v.3

Graduate student (law), 25. "Reading room with a window. In the window I could see the room behind me and, in it, standing very tall by the door, someone whose outline did not match anything in the room. He had his back half to me. When I moved, the angle never improved. Whoever he was, he was being careful of the glass."

Coder note (A.F., 13 DEC 2006): Figure in window reflection, facing slightly away. Matches S-025.

Convergence flag: H2 · R2.

LOG-1157 · S-058 · 12 DEC 2006 · Night 17 of 20 · Consent v.3

Ferry ticket clerk, 41. "I've kept every log in full, as asked, even the boring ones. Tonight a clerk at the counter was reading my previous entries aloud to someone behind him, ticking off lines. When he got to the sentence about the green coat he stopped and said, quite gently, that it was a little more grey. I woke annoyed."

[Second hand, blue-black ink, beside "green": "grey."]

Coder note (N.H-L., 05 JAN 2007): Correction appears on the original form. Coat in the participant's wardrobe is grey. Participant reports having "always said green."

Convergence flag: A2 (annotation, accurate).

LOG-1200 · S-060 · 18 DEC 2006 · Night 20 of 20 · Consent v.3

Retired surveyor, 68 (profession noted without comment). "Last night of the study. They asked me to sign for my coat. I said I'd signed for it already, a long time ago. The Usher agreed. He said that was why they were asking. I woke at 3:07. The bed was a hand's width closer to the window than I'd left it."

Coder note (A.F., 02 MAY 2007): Signature request ("sign for my coat"). HS not described. Occupation entered as recorded.

Convergence flag: S2 · D2.

Cohort F · S-061 to S-072 · 19 SEP 2007 to 06 NOV 2007 · 4 logs shown

LOG-1305 · S-066 · 19 SEP 2007 · Night 5 of 20 · Consent v.4

Janitor (secondary school), 58. "Chairs in rows, nine, grey. I've stacked chairs for thirty years and these were wrong somehow, spaced for people I couldn't see. A tall figure was reflected in the glass of a fire-notice frame. He did not look like the same figure as in the stairwell, but then I only saw him there in the window, in the dark of the glass."

Coder note (A.F., 07 OCT 2007): Nine chairs; reflected figure. Overnight technician's shift log for this date has no entry after 03:07.

Convergence flag: B3 · H2.

LOG-1358 · S-068 · 02 OCT 2007 · Night 18 of 20 · Consent v.4

Jeweller, 62. "Went down this time, as far as the door with the B on it. A man in a grey waistcoat there asked how I'd come to be there. I gave him my first answer, which was that I'd been sent. He declined it. Politely. He asked for another. I said I was lost. He accepted that one and wrote 'lost' in the book."

Coder note (B.O-S., 19 OCT 2007): "Declined my first answer." Cf. Subject 0 file. Participant did not go through the B door.

Convergence flag: F2 · A0.

LOG-1253 · S-063 · 22 OCT 2007 · Night 13 of 20 · Consent: none on file

Postal worker, 45. "At the counter, a clerk laid out items from my kitchen, one at a time, as if for valuation. A mug. A bread knife. The tea towel. He asked whether each was accurate. I said yes, all of it. He wrote it down and said that would be all. Back home the mug stood on the shelf a little to the left of the ring it had left."

Coder note (N.H-L., 08 NOV 2007): Clerk exhibits participant's property. Mug displacement 3 cm, participant-measured. Consent form: see administrative note.

Convergence flag: B3 · D1.

LOG-1407 · S-071 · 06 NOV 2007 · Night 7 of 20 · Consent v.4

Court translator, 40. "A long corridor, wood floor, doors both sides with frosted glass. At the end, a mirror, floor to ceiling, and something standing in it, between two lamps, not in front of them. It was tall and narrow, like a coat rack with an opinion. I walked towards it. It did not approach. Nobody in that building has ever once run."

Coder note (A.F., 04 JAN 2008): Mirror figure. Different build and outline from S-025's report; same stance (facing away).

Convergence flag: H3 · R3.

Cohort G · S-073 to S-084 · 12 MAR 2008 to 22 APR 2008 · 4 logs shown

LOG-1479 · S-074 · 12 MAR 2008 · Night 19 of 20 · Consent v.4

Retired bus conductor, 72. "I counted the stairs the way you do with a bus, out loud. Nineteen down. Coming up, eighteen. I'd have liked to have it the other way round. The man at the top said it never is. He had a card with the numbers already written on it. Mine were right."

[Second hand, red pencil, bottom of page: "Count confirmed. Thank you, S-074."]

Coder note (B.O-S., 02 APR 2008): Red-pencil note is in the log keeper's style but she states she did not make it. Verify against LOG-0752.

Convergence flag: C3 · A3.

LOG-1614 · S-081 · 03 APR 2008 · Night 14 of 20 · Consent v.4

Night pharmacy assistant, 34. "Waiting room, brass bell, board at 40. I filled in the log last night exactly as asked, not leaving anything out. The clerk at the counter had my entry in front of him with one word crossed out and another put above it. He read the new word aloud. It was a better word than mine. I thanked him. He said I was welcome."

Coder note (N.H-L., 02 MAY 2008): Participant describes receiving a correction in the dream, then finds the same word replaced on the physical log. HS: none.

Convergence flag: B3 · A3.

LOG-1528 · S-077 · 08 APR 2008 · Night 8 of 20 · Consent v.4

Doctoral student (hydrology), 27. "A hook with my coat on it, as before. This time there was a label on it with my name, in my handwriting, but I don't write my 'g' like that. I took the label off. I woke with the coat on the chair, labelled. The label was written the way I do it. I don't know which one of us corrected it."

Coder note (A.F., 29 APR 2008): Label first in "wrong" hand, then corrected. Coat hook; cf. LOG-0708.

Convergence flag: B3 · A2.

LOG-1680 · S-084 · 22 APR 2008 · Night 20 of 20 · Consent v.4

Retired locksmith, 66. "Last night. Waiting room, board at 40. The clerk said that was all. He stamped something and slid it across: 'Reviewed.' He said if I was ever asked to sign, I should say I'd already been seen. I asked what happens after. He said, filing. I woke at 3:07. The room was a little narrower. I've measured."

[Second hand, blue-black ink, below signature: "Reviewed."]

Coder note (B.O-S., 12 MAY 2008): Final participant, final night. Wording "Reviewed" appears in LOG-1681 (PI). Compare also S-043, S-060.

Convergence flag: B3 · A3 · S3.

Submit a log

Visitors may submit a log entry for review. Entries are reviewed by the retention office.

Date	Time in	Time out	Badge	Initials
28 OCT 2006	22:50	07:05	S-052	PP
29 OCT 2006	22:48	07:10	S-052	PP
30 OCT 2006	23:41	—	HR-0314	EH
30 OCT 2006	03:07
31 OCT 2006	—	06:52	HR-0314	EH

Location	Item	Register	Found
Room 1	Bed, hospital type	1	1
Room 1	Chair, grey, vinyl	2	2
Room 2	Bed, hospital type	1	1
Room 2	Chair, grey, vinyl	2	2
Room 3	Bed, hospital type	1	1
Room 3	Chair, grey, vinyl	2	2
Room 4	Bed, hospital type	1	1
Room 4	Chair, grey, vinyl	2	2
Waiting area	Chair, grey, vinyl	6	7