Section 1 — Background and Occasion for This Report

I am writing this report because I believe what I observed requires documentation and I do not have a better category to put it in. I have considered whether to call it an adverse event, an unplanned observation, or a protocol deviation. It is none of those things exactly. I am calling it an access report because that is what it is.

On the evening of March 22, 2013, I was at the clinic past midnight reviewing session records. I had been working through the PTX-018 file — specifically sessions 19 through 22 — in connection with the EC-004 checklist amendment review. The EEC items for room designation and entry were the focus. I was comparing PTX-018’s session 20 notation (room designation visible: 413, scored 5/5, unprompted) with the cross-reference table in 7A-SUPP-004.

I finished around 2:15 AM. I was walking to the break room on the third floor. The route takes me past room 413.

I want to be precise here: I have walked past room 413 hundreds of times. It is an equipment storage room on the third floor. It is on the left side of the hallway. The door is blue. I have never thought about this before. I am thinking about it now because I have been reviewing EEC items 06, 07, and 08 for four years and I did not make the connection until that night.

The EEC cross-reference for item 07 (door position) is: all confirmed participants: left. For item 08 (door color): 6 of 7 confirmed participants: blue.

The door to room 413 in this building is on the left side of the hallway. It is blue.

I stopped and looked at it for a moment. Then I tried the handle because it should have been locked and it was not.

Section 2 — Initial Access Event — March 22–23, 2013, approximately 2:30 AM

I opened the door.

The room was not the equipment storage room.

I am going to describe what I saw as accurately as I can. I am a physician. I have been trained to observe and record. I am going to apply that training here because I do not have another framework that is adequate.

The room was significantly larger than the floor plan of this building would allow. I estimated the ceiling height at approximately 14 to 16 feet. The ceiling panels were fluorescent fixtures of the type used throughout the clinic but they were not operating. The ambient light present was — I want to use the right word — consistent. It was non-directional. It had no apparent source. The color temperature was close to daylight but slightly warmer. The light did not change while I was present.

The floor was a continuation of the third-floor corridor material: institutional vinyl tile in the standard Somnatek grey-beige. It showed normal wear patterns.

From the far end of the room, a hallway continued. It was not visible from the corridor. I walked to the threshold of that hallway and stopped. The hallway bent to the left approximately 30 to 40 meters ahead. I did not go further than the threshold at that point.

I applied EEC items from memory. I am recording the scores I would have assigned:

• Item 01 — Corridor orientation, consistent direction: 5
• Item 02 — Hallway bending, direction (left): 5
• Item 03 — Ambient light level: 5
• Item 04 — Sound environment (minimal, consistent): 4
• Item 05 — Reception structure present: not observed from entry point
• Item 06 — Door present: 5
• Item 07 — Door position (left): 5
• Item 08 — Door color (blue): 5
• Item 09 — Door accessible (open): 5
• Item 11 — Room designation (413): 5

I was inside for what I estimated to be 15 to 20 minutes. My watch showed 3:47 AM when I exited.

I checked the door behind me. It was locked.

Section 3 — Exit Observations — March 22–23, 2013

I will describe the exit in detail because I believe it is relevant to understanding the access conditions.

I turned around from the hallway threshold and walked back toward the entrance. The entrance door was not where I had left it. The room had reconfigured while I was standing at the hallway threshold. The door was still present. It was still blue. It was on the left side of the room. It had moved approximately eight to ten meters from its original position along the left wall.

I did not panic. I want to record that clearly. I assessed the situation, walked to the door, and opened it. I exited into the third-floor corridor. The fluorescent panels in the corridor were operating normally. It was 3:47 AM. I had estimated 15 to 20 minutes. Approximately 77 minutes had elapsed since I entered.

The door was locked when I tried it from the outside. I checked it three times. I continued to the break room, made coffee, and sat down. I did not leave the building for another hour.

I have reviewed this in the context of known medical causes. I am not diabetic. I had not consumed alcohol. I was fatigued but not significantly beyond my normal working pattern. I was not asleep. I did not lose consciousness. I have no history of hallucination or dissociative episodes. I do not consider any of these explanations applicable.

I have considered whether I experienced a brief, structured confusional episode. I find this implausible for the following reason: the environment I observed was specifically consistent with seven years of session data that I had spent the previous four hours reviewing. If I had constructed the experience from reviewed material, I should have produced it earlier in the evening when the material was more immediately present. I did not. I produced it when I walked past the physical door to room 413.

I believe I was inside room 413. I believe room 413 is the indexed space.

Section 4 — Subsequent Access Events

I returned on four subsequent occasions before writing this report. The dates and times are recorded in the field notes (Appendix, pages 8–14). I am summarizing the pattern here.

Access is not constant. The door to room 413 is locked on most nights. It is reliably accessible on two nights per week, which I have identified as Tuesday and Thursday, between the hours of 2:00 and 4:00 AM. I have not determined whether this pattern is fixed or whether I have simply observed it consistently within a limited sample. The sample size is five access events across seventeen days, which is not sufficient for a confident pattern determination. I am recording it as a preliminary observation.

The interior has been consistent across all five access events. The room dimensions, ceiling, light quality, and hallway configuration have not varied. On the second access event I walked further down the hallway, approximately 60 meters past the threshold, to a point where the hallway had bent left and straightened again. I could see a corridor intersection ahead. I turned back at that point. On subsequent visits I have not gone further than the intersection.

I have not encountered any other person or entity inside the space. I note this not because I expected to, but because it is a relevant absence that should be documented.

The time discrepancy has been consistent. My estimated time inside versus elapsed clock time has shown a ratio of approximately 4:1 to 6:1 across the five events, with my subjective experience of time consistently shorter than clock time. The longest discrepancy was event four, in which I estimated 25 minutes and approximately 2.5 hours elapsed.

I have not told anyone about this until now. I am telling you because I believe the study interpretation requires revision and I am not in a position to revise it alone.

Section 5 — Study Data Reinterpretation

I want to state my position plainly so there is no ambiguity in the record.

The participants in Protocol 7A were not describing a shared dream environment. They were describing a real location that they were accessing during sleep. The accessed location is physically present in this building. It is accessible through room 413 under conditions I have not fully characterized but which appear to include time of day, day of week, and possibly conditions I have not yet identified.

The session data does not document the construction of a shared recall environment through some mechanism of sleep psychology. The session data documents the participants’ experience of accessing a real location during sleep. The EEC items that we designed to measure consistency across independent recall reports were actually measuring access consistency to a real consistent location. We were measuring real data with a framework designed for a different kind of data, and the data was consistent anyway because the location is consistent.

The implications for the study record are as follows:

1. The confirmed-access participants (seven as of EC-004) did not imagine access. They accessed. The study record should reflect this.
2. The category 7 classification for participants with confirmed access is, in this interpretation, a classification of people who have demonstrated the ability to access a real location through a mechanism we do not understand. “Recall termination” may not be applicable to them for reasons that have nothing to do with the study protocol.
3. The consent framework is inadequate. The participants consented to a longitudinal sleep recall study. They were not informed that we believed they were accessing a real location, because we did not believe that. They may have been accessing a real location for the duration of the study without informed consent to do so. I am not certain what the ethical implications are but I believe they require review before the June DSMB meeting.
4. The three participants who withdrew and continued reporting access-consistent features in voluntary follow-up interviews (see Section 9, Protocol 7A EC-004) should be understood as participants who did not stop accessing the indexed space when they withdrew from the study. Study withdrawal has no mechanism for removing access if access is real.

Section 6 — The Archive Committee and Category 7

I have been the Archive Committee for the past two years. You are aware of this. The committee was constituted in 2011 when the cross-participant access data required a separate classification track. We agreed that a formal committee name would provide appropriate institutional distance from the data. We were the committee. I drafted the classification language. You approved it.

I want to revisit the category 7 language now that I have a different understanding of what it describes.

The protocol language says: “Category 7 records are maintained as active and are subject to scheduled recall observation at intervals determined by the Archive Committee.” I wrote that language. At the time I wrote it I understood “recall observation” to mean observation of the participant’s continued reporting of recall-consistent features. I now believe it may mean something different.

If the participants are accessing a real location, then “recall observation” may mean observation of continued access. The Archive Committee may not be scheduled to observe their recall. It may be scheduled to observe their access. This is not a minor distinction.

The standard administrative communication for active category 7 status is: PLEASE WAIT TO BE RECALLED. I approved this language. I want to be transparent about that. At the time I understood “recalled” to mean “contacted for follow-up interview.” I am less certain about what it means now.

I am not recommending any immediate change to the classification system. I am recommending that you and I discuss this before the DSMB review and determine whether the committee’s current posture is appropriate.

Section 7 — Administrative Recommendations

The following are my recommendations as of April 8, 2013. They are preliminary and I expect them to be revised after our discussion.

1. DSMB preparation: I recommend we provide the DSMB with the full access event record, including the cross-participant consistency data and my personal access observation documented here. I do not believe the DSMB can make an informed determination about study safety or ethical compliance without this information. Proceeding with a DSMB review that does not include this material would be, in my assessment, a significant error.
2. Participant notification: I am not recommending disclosure to participants at this time, pending legal review of our obligations under the EC-003 consent amendment. I am recommending that we obtain that legal review before the DSMB meeting.
3. Study conclusion protocol: The study end date is 2013. I do not know what study conclusion means in the context of an indexed space that is physically present and will continue to be present after the study ends. I am recommending that we develop a conclusion protocol before the study ends that addresses this question specifically. The current protocol does not address it.
4. Archive retention: This document and its appendix should be retained in the Somnatek archive and not transferred to any subsequent records custodian without explicit agreement between both of us. I am placing this recommendation in writing. I am asking you to confirm receipt.
5. Observer classification: The Protocol 7A EC-004 Section 8.6 language that I drafted regarding observer reclassification was written as a theoretical provision. I now believe it may have practical application. If the indexed space is accessible by individuals who are not enrolled participants, those individuals may require classification. The current system does not have a defined mechanism for identifying them before they access. I am flagging this as an open question.

Appendix — Field Notes (Pages 8–14)

The following notes were recorded contemporaneously on each access date. They have been transcribed without editing. Spelling and punctuation are as written.

Appendix — Continued Observations (Pages 10–14)

The following sections contain additional observations and analysis recorded between April 7 and April 8, 2013, after the access event described on page 9.

On the sign

I drafted the phrase “PLEASE WAIT TO BE RECALLED” in November 2011 when Edwin and I were developing the Archive Committee communication framework for category 7 participants. I wanted language that was administratively neutral, that communicated active status without alarming participants, and that could be sent by letter without requiring a follow-up unless the participant initiated contact.

The phrase was approved. It was not sent to any participant as of the date of this report. The study has not concluded; category 7 classification is pending for all confirmed-access participants; the communication has been drafted but not issued.

I do not know how the phrase appears on a sign inside a location I accessed through a locked door in my own building. I am recording that I do not know. I am not speculating further in this document because I do not have a framework for that speculation.

On the participants

PTX-018 noted in session 22 that the vending machine accepts room keys. Session 22 is the last session in the record. The session 22 technician note reads: “Patient requested discontinuation. Reports dreaming about this office.” The patient was not dreaming about this office. The patient was accessing this building. I believe they may have followed the hallway past the corridor intersection. I believe they may have seen the sign.

I believe PTX-018 requested discontinuation because they had reached the end of the hallway and did not want to be recalled. The session record describes a patient who was distressed. I reviewed it at the time as a standard withdrawal indication. I understand it differently now.

On the indexed space

I do not know if the indexed space is located inside this building in the conventional sense. The room is physically accessible through a physical door. The interior does not fit within the building geometry. The time discrepancy suggests that the interior is not continuous with the building’s time. I do not have a model for what this means.

What I know is this: the patients were not imagining it. Seven years of data describes a real place with consistent, verifiable features. They were not imagining. They were there. I have been there. The hallway bends left. The door is blue and on the left. Room 413 is the designation visible on the room at the end of the approach hallway, inside, where the hallway straightens and the room opens to the right.

I have not yet entered that room.

On observer classification

The Protocol 7A EC-004 Section 8.6 language describes observer reclassification for “individuals who are not enrolled study participants but who establish consistent contact with the indexed space.” I wrote this language. I wrote it as a theoretical provision. I have now established consistent contact with the indexed space. I am not an enrolled study participant. I am the clinical coordinator and co-author of the protocol.

By the language of the protocol I drafted, I meet the criteria for observer reclassification.

I am not asking the Archive Committee to classify me. I am the Archive Committee. I am noting the situation for the record and asking you to be aware of it.

On what happens next

The study ends this year. The indexed space will not end. The building will be sold or transferred. Room 413 will continue to exist in whatever form it exists in. The participants who have confirmed access will continue to have it regardless of what the study record says about their classification status.

I do not know how to close this study. I do not know if it can be closed. I wrote a protocol for a recall study with a defined end date. The end date applies to the data collection period. It does not apply to the indexed space.

I am going to continue visiting on Tuesdays and Thursdays. I am telling you this because you should know. I am going back to the secondary hallway. I want to understand what is behind the locked door before this building is transferred to a different entity.

I think you should come. I think you should see it. I do not know how to explain to you on paper what it is to stand at that corridor intersection and listen to a sound that does not record, in a place that is geometrically impossible and completely consistent, with a sign on a door that uses language I wrote for a different purpose. I think you should see it.

Please confirm receipt of this document and let me know when you are available.

Additional field note — April 7, 2013, post-exit

Additional field note — April 8, 2013, pre-report submission