Lucid episodes in people with advanced dementia: characterizing caregiver reports to advance definition and measurement of episodes

Abstract

Background and Objectives

Lucid episodes (LEs) in advanced dementia are characterized by spontaneous, transient recovery of functional abilities. Despite recent increases in research on this previously understudied phenomenon, standard definitional criteria are lacking, presenting challenges for research. Researchers have proposed a preliminary framework to guide uniform labeling and operationalization of salient data elements for investigating LEs and developing definitional criteria. This study leverages this framework to characterize reports of LEs from caregivers and advance their definition and measurement.

Research Design and Methods

Family caregivers (N = 259) completed an electronic survey with quantitative and open-ended questions about LEs they witnessed (N = 479 episodes). Respondents were purposively sampled for qualitative interviews. Open-ended question responses (n = 208) and in-depth interviews (n = 24) were analyzed using content analysis.

Results

Caregiver descriptions suggest that the content and context of LEs are heterogeneous. Descriptions included observable changes from baseline, with behavior or communication that was distinguishable from routine cognitive fluctuations. LEs were sometimes, but not always specified by logically organized actions; recognition or awareness of people, emotions, and surroundings; and insight of their cognitive limitations. Many LEs required deep personal knowledge to attribute it as a meaningful shift.

Discussion and Implications

Findings suggest that definitional features may universally specify a significant return of ability with more flexible criteria for how this return manifests in observed actions and domains of recovery (i.e., insight, function, and communication). The impact of LEs on observers may be better measured separately, although further research should explore this.

Background and Objectives

Current models of Alzheimer’s disease and related dementias (ADRD) assume that lost cognitive capacity is irretrievable. However, numerous case studies, anecdotes, and now prospectively documented reports describe instances when people living with ADRD exhibit transient recovery of abilities and mental clarity in late stages of the syndrome (Batthyány & Greyson, 2021; Benson et al., 2023; Bostanciklioğlu, 2021; Mashour et al., 2019; Nahm et al., 2012; Ramirez et al., 2024). These lucid episodes (LEs) are sometimes referred to as paradoxical lucidity (Eldadah et al., 2019) because of their inconsistency with the prevalent ADRD neuropathology model and lack of scientific explanation (Mashour et al., 2019; Nahm, 2022) and are thought to be markedly different from routine cognitive fluctuations. Until recently, LEs have received relatively little scientific attention, and therefore, researchers’ understanding of them is limited (Eldadah et al., 2019). Despite the growing body of research on LEs, there is still no scientific consensus on a precise definition of LEs (Bangerter et al., 2024). Recent concept analyses and critiques about the phenomenon have highlighted the need to clarify core attributes and indicators of LEs—such as cognitive clarity, abilities or awareness, and meta-awareness, but have also cautioned against using overly rigid definitional criteria that may exclude poorly understood or overlooked indicators or introduce bias in the interpretation of the phenomenon at this early stage of inquiry (Bostanciklioğlu, 2021; Eldadah et al., 2019; Mashour et al., 2019; Morris & Bulman, 2020; Nahm, 2022; Peterson, Clapp, Largent, et al., 2022; Ramirez et al., 2023). Thus, there is a critical need to conceptualize and characterize LEs using scientific methods that can advance precision in definitions and measurement. Gaining a qualitative understanding of the breadth of LE experiences is paramount, as it reveals the potentially nuanced ways that an indicator may be present. This depth also ensures that measures of LE indicators can be measured with validity and reliability.

Given the early state of scientific inquiry and definition development, Gilmore-Bykovskyi et al. (2023) proposed a Preliminary Research Framework for Lucidity (PRF-L) for researchers to consider when operationalizing and reporting research that characterizes LEs. The PRF-L includes five questions intended to guide researchers toward refining evidence-based definitions and advancing consistent measurement of LE attributes across studies (Gilmore-Bykovskyi et al., 2021). A summary and explanation of the questions is as follows:

1. Do LEs occur across the continuum of the ADRD trajectory? This question addresses whether LEs are associated with end-of-life or only within the latest stages of ADRD; or if they also extend more broadly across the ADRD continuum and potentially share mechanisms with cognitive fluctuations in earlier stages.

2. Are LEs all individually defined events across ADRD or do they have uniform qualities? This question addresses whether a set of common attributes are present in all LEs and are, therefore, generalizable, or if each LE can only be defined based on the relevant change within an individual.

3. Do LEs include meta-awareness? Is insight about one’s cognitive impairment a distinct attribute of lucidity? This question addresses whether meta-awareness, or an individuals’ ability to recognize and be aware of their own cognitive deficits despite their ADRD, is a common attribute of LEs. Meta-awareness is sometimes referred to as “insight.”

4. Are speech, language, and communication measures necessary and/or sufficient to identify the presence of a LE? This question captures two different concepts. First, it addresses whether forms of communication other than speech, such as nonverbal actions or movements and utterances, and quality of communication, such as coherence of speech and pacing, are relevant for identifying LEs. Second, it addresses the conceptual and methodological question of whether the form and quality of the communication in a LE can be identified directly (e.g., through video or audio recording), or whether appraisal from someone who understands the cognitive shift and personal context of the communication is necessary.

5. Do LEs have intrinsic temporal markers or a necessary magnitude of change in order to be recognized? This question poses whether specific thresholds must be present to be considered a LE (e.g., frequency, duration, antecedents, and consequences) and whether a certain magnitude of recovery in abilities is necessary to specify an event as an LE.

Qualitative data are often highly valuable for advancing from conceptualizing a phenomenon to operationalizing it using valid and reliable measurement (Carpenter, 2018). To contribute to the conceptual and operational debate about lucidity, we applied Gilmore-Bykovskyi’s PRF-L (Gilmore-Bykovskyi et al., 2023) to a robust set of qualitative data from current and former family ADRD caregivers who reported witnessing LEs. The aim of this study is to further describe potential LE indicators and their complexities using this framework to help identify candidate items for measurement and elucidate potential biases and strengths for measurement strategies.

Methods

This study used a descriptive qualitative design in order to better understand potential definitional criteria for LEs and explore important conceptual and methodological considerations for future research of the phenomenon (Kim et al., 2017; Neergaard et al., 2009; Sandelowski, 2000). This qualitative design is ideal when the research questions are straightforward, data explicitly describe something that has happened, and the need for interpretation is minimal beyond the description of the phenomenon (Sandelowski, 2000; Sandelowski, 2010).

The researchers initially surveyed current and former family caregivers of people living with ADRD to inquire whether they witnessed an LE. The survey included a general description of LEs, which is based on working definitions developed by the National Institutes on Aging: “A lucid experience is unexpected, spontaneous, meaningful, and relevant communication with your relative, friend, or neighbor when they had lost the ability to speak or have personal interactions” (Eldadah et al., 2019). If a caregiver had witnessed an LE, they were asked to describe the episode(s) quantitatively. Using the quantitative data, the researchers developed a typology, which included four types of LEs (Griffin et al., 2024). Also included in the survey was an open-ended question that captured qualitative caregiver appraisals of the reported LEs (“In your own words, explain what you observed. Feel free to share what your relative, friend or neighbor talked about, gestures they made, or requests they had during their lucid experience.”). The researchers then purposively sampled those reporting they witnessed an LE to participate in qualitative interviews, which enabled an in-depth exploration of the circumstances and features of the LE. Data for this study were drawn from both the open-ended responses in the survey and the qualitative interviews.

Sample

Caregivers were recruited from UsAgainstAlzheimer’s, an advocacy group that aims to mobilize participation in research on effective treatments and quality care for ADRD (usagainstalzheimers.org). UsAgainstAlzheimer’s administers the A-LIST, a unique online community of individuals who self-identify either as someone at risk for or with mild cognitive impairment or ADRD; a current or former caregiver for someone with ADRD; or, someone interested in brain health (alist4research.org).

Survey procedures

In February 2021, UsAgainstAlzheimer’s sent email invitations to A-LIST participants (N = 3,577; see Figure 1 for the sample selection). Invitations requested participation from current and former caregivers of someone living with ADRD, described the study’s purpose and a general description of a LE (i.e., “… an unexpected change in his/her thinking or concentration, where he/she has been lucid, alert, aware of his/her surroundings, or able to communicate in meaningful ways. This could be verbal or non-verbal communication.”), and provided a hyperlink to an electronic survey. Non-responders were sent a reminder 4 and 15 days after the original invitation. UsAgainstAlzheimer’s Institutional Review Board (IRB) approved the survey study, and all caregivers provided informed consent prior to participating in the survey. Deidentified survey data were shared with the study team.

Figure 1.

Flow chart of the sample selection. MCI = Mild Cognitive Impairment.

Caregivers were asked if they had ever witnessed an LE in a relative, friend, or neighbor who has ADRD, and if so, characteristics of the person who experienced the LE, including a question on cognitive status on a typical day before the LE with seven ordinal responses options (1 = unconscious; 2 = mostly asleep; 3 = mostly awake, but not responding or reacting; 4 = extreme difficulty with memory, attention, or focus; 5 = moderate difficulty with memory, attention, or focus; 6 = minor difficulty with memory, attention, or focus; 7 = fully aware, no impairments). With lucidity expected to occur in later stages of ADRD, the researchers limited their analyses only to those reporting options 1–5.

Interview procedures

Survey respondents were also asked if they would be willing to participate in a semi-structured interview with a trained qualitative researcher who had no personal experience of witnessing LEs. Participants were informed that the interviewer would be from USAgainstAlzheimer’s research partner, Mayo Clinic, and a new consent for the interview would need to be provided to that partner. With the sample of those interested in ongoing research participation, the researchers populated strata from which they could sample. Strata included the type of LE based from the typology analysis (Griffin et al., 2024) and by their relationship to the care recipient (spouse vs. adult child). Within each stratum, they then considered sampling by living arrangements (e.g., caregivers coresiding with the person living with ADRD at the time of the LE), race/ethnicity, and gender of caregivers and the person living with ADRD to capture varied experiences.

Sample size was based on Malterud’s concept of information power (Malterud et al., 2016), which asserts that the more information a sample has about a phenomenon, the fewer participants are required. Knowing that everyone had experienced an LE from the survey, the researchers estimated they would need at least 24 people (i.e., three people per LE type/relationship cell). After this initial sample was interviewed, the researchers reviewed data and determined saturation had been achieved. They concluded no additional participants were needed.

Questions for the interview guide were informed by survey results and reviewed by the study’s caregiver advisory committee—a committee made up of a racially and ethnically diverse group of former caregivers of people living with ADRD who provided guidance on appropriate and respectful ways of engaging caregiver participants, survey development, and communication with caregivers. Interview questions included the history of the person living with ADRD’s disease progression, descriptions about the LE, the circumstances preceding and following the episode, and caregiver reactions to the LE (see Supplementary Figure 1 for the interview guide). The Mayo Clinic IRB approved the protocol for the semi-structured interviews, and those interviewed provided verbal consent. Interviews were conducted over the telephone by an experienced qualitative analyst (DF) on the study team. Interviews were audio-recorded and transcribed verbatim by a research assistant, and deidentified.

Data analysis

Answers to the open-ended survey question (OSQ; n = 208) and transcripts from interview recordings (IR; n = 24) were coded and analyzed using conventional content analysis methods (Kim et al., 2017; Patton, 2014). Illustrative quotes in the results are labeled as OSQ and IR to indicate their source.

Two qualitative researchers (DF and JMG) met throughout the analytic process to ground themselves in the data and the guiding questions. The analysts reviewed the guiding questions and identified key words, methods, and concepts associated with each question (e.g., meta-awareness, insight, and self-aware) as a first set of codes. Additional codes were generated by reading transcripts line-by-line. Memos were documented with the researchers’ assumptions and initial thought processes about coding and then tagged data with codes. In a second stage of analysis, codes were refined and mapped to correspond with the five conceptual questions in the PRF-L from Gilmore-Bykovskyi et al. (2023). In a third stage of analysis, PRF-L codes were compared to the initial codes that did not map onto the PRF-L codes to determine other potential conceptual or methodological descriptions or attributes of LEs. These were then coded inductively. Coded data were also organized and analyzed by the level of awareness or cognition prior to the LE, as reported in the survey. A final codebook was developed and those codes were applied to each transcript by DF and JMG. To facilitate analysis and text data queries, data were entered into NVivo software (NVivo 13, QSR International Pty Ltd). Differences in coding were resolved through discussion. The full research team then met to review the analysis, codebook, and results, and provided additional feedback to assure there was consensus about the findings.

Results

Survey invitations were sent to 3,577 people, and 538 responded to the request (see Figure 1). Of those, 58 were not eligible (i.e., not current or former caregivers). Of the 480 eligible caregivers, 294 (61.1%) reported witnessing an LE. In the survey, 259 caregivers provided detailed information about at least one LE (total 479 episodes), 208 answered the open-ended question describing an episode, and 179 agreed to follow-up for additional research. Twenty-seven of the 179 who provided survey responses about an episode witnessed were invited for in-depth interviews, and 24 completed interviews. Interviews lasted an average of 33 minutes (range = 14–63). Table 1 shows the demographics for caregivers of the open-ended questions and those interviewed and caregiver perceptions of the care recipient’s awareness prior to the LE.

Table 1.

Characteristics for caregiver participants and their care recipients.

Variables	Interviews, n (%)(N = 24)		Open-ended survey questions, n (%) (N = 208)
Caregiver sex (% female)	21	(88)	142	(68)
Caregiver age
31–40	0	(0)	2	(1)
41–50	1	(4)	12	(6)
51–60	5	(21)	44	(21)
61–70	13	(54)	80	(38)
71–80	5	(21)	57	(27)
81+	0	(0)	12	(6)
Unknown	0	(0)	1	(1)
Caregiver’s highest level of education
High school and under	4	(16)	21	(10)
Associate’s degree	7	(29)	29	(14)
Vocational degree	3	(13)	24	(12)
Bachelor’s degree	3	(13)	62	(30)
Master’s degree	6	(25)	61	(29)
Doctoral degree and above	1	(4)	9	(4)
Caregiver’s race
White	16	(67)	189	(91)
Black/African American	5	(21)	10	(5)
Native American/Alaska Native	2	(8)	2	(8)
Other races	1	(4)	7	(3)
Caregiver’s Hispanic ethnicity	2	(8)	13	(6)
Caregiver’s relationship to care recipient
Spouse/partner	7	(29)	138	(66)
Adult child	15	(63)	58	(28)
Other (e.g., sibling, friend, neighbor)	2	(8)	12	(6)
Caregiver co-residing with the care recipient	8	(33)	84	(40)
Type of dementia
Alzheimer’s disease	17	(71)	129	(62)
Lewy body	2	(8)	13	(6)
Parkinson’s	1	(4)	24	(12)
Frontotemporal	0	(0)	17	(8)
Vascular	2	(8)	22	(11)
No diagnosis/don’t know	2	(8)	3	(1)
Care recipient still living at time of survey	5	(25)	73	(35)
Care recipient’s level of awareness prior to LEs
Fully aware with no impairments	1	(4)	4	(2)
Minor difficulty with memory, attention, or focus	1	(4)	3	(2)
Moderate difficulty with memory, attention, or focus	4	(17)	41	(20)
Extreme difficulty with memory, attention, or focus	10	(42)	95	(46)
Most or much of the time awake but not responding or reacting to anything around them	2	(8)	28	(14)
Mostly asleep, but when awake, not aware of surroundings	2	(8)	14	(7)
Unconscious	0	(0)	0	(0)
Missing	4	(17)	23	(11)

Note. LE = lucid episode.

Caregivers described LEs as indelible moments and used vivid detail, including the time, place, and circumstances surrounding the LE. Specifically, caregivers provided information on the level of cognition prior to the episode, communication content, patterns and structure, and why they considered it to be a LE. Therefore, quotes were often coded for multiple themes.

Guiding question 1: Do LEs occur across the continuum of the ADRD trajectory?

Caregivers described LEs across the ADRD trajectory. Consistent with other studies and case reports (Nahm et al., 2012), descriptions of LEs later in the disease course and close to death were often reported as especially remarkable. One caregiver who reported her mother had extreme difficulty with memory, attention, and focus prior to the LE reported: “The day she died, when we came to see her… she took off her oxygen mask and said, ‘Well, hello, Joe’ to my husband. This was after months of not using our names and not making much sense.” (OSQ #185). Similarly, a caregiver described her parent’s recovery of verbal abilities before dying: “… the final lucid experience was the day before he died. I came over to his home and he was rarely awake and never spoke at this time. But we talked and talked to him regardless. I went over to his bed, he was awake, and I leaned into his view and said, ‘Hi Dad!’ and clear as a bell he looked at me and said, ‘Hi Honey!’ That’s the last he spoke” (OSQ #7).

Other LEs occurred in people with advanced disease, but years before death. For example, one caregiver reported that prior to the LE, her parent’s routine state at the nursing home where she lived was mostly awake but not responsive: “We brought with us her favorite cat and after having her recline on the bed we put the cat on her in a way that was familiar to her. After petting her cat for a minute my mom called her ‘baby.’ Even though she was previously unable to speak or even say the names of her children… she knew the pet name of her cat and said it clearly. This happened about 3 years before she passed away. The rest of the times she babbled with few coherent words” (OSQ #10).

There were also reports of LEs from people who were described as having moderate stage dementia, but had exhibited what respondents felt was short-term recovery of more cognitively complex abilities than they had previously been able to perform. For example, one caregiver who reported his father having moderate difficulty with memory, attention, and focus prior to the LE, recalled: “My father, after being admitted to a SNF [skilled nursing facility], found shaving cream and wrote ‘Help Me’ backwards on the window so it could be read from the outside” (OSQ #108). Another caregiver recalled her parent, who reportedly had moderate cognitive impairment: “… expressed remorse for the stress I must be feeling in caring for him alone.” (OSQ #66).

Guiding question 2: Are LEs all individually-defined events across ADRD or do they have uniform qualities?

Questions persist in the literature about how best to characterize cognitive fluctuation across the course of dementia (Batthyány & Greyson, 2021). Most caregivers described lucidity in terms of noticeable shifts in the clarity and coherence of verbal or nonverbal communication or abilities which were sometimes very brief and sometimes longer, but were all different from routine cognitive fluctuations. For example, one caregiver reports: “A week before she died my mother received the Catholic ‘last rites,’ a special blessing and anointment. Her mind was very far gone at his time and she was physically very weak. But she looked up at the priest, smiled and very clearly said ‘Thank you.’ My father and I were both present and believed she knew she’d been blessed” (OSQ #9).

Codes mapped to this question also included descriptions of a diverse range of abilities and events in which the descriptions of the recovered abilities were highly specific to the individual and required a familiarity of the context or personality of the person living with ADRD to be understood or considered unusual. One caregiver, for example, described details known only to immediate family members: “[My husband] had always wanted a boy (we have 2 daughters). He nicknamed our older daughter ‘Pat.’ That stuck with our daughter from age 3 to late teens. One day our daughter and I were visiting [my husband]. [He] was unusually quiet. As our daughter was leaving, she said ‘bye Dad, see you later.’ She turned and started to leave: [my husband] said, ‘bye Pat.’ He hadn’t used that nickname in years! We both heard it very distinctly and both stopped short in our conversation” (OSQ #36). Another caregiver reported an LE with her sister who she described as having extreme cognitive impairment: “My sister [the person living with ADRD] was a speech pathologist. I was telling her about my grandson’s speech issue and she explained the problem and how it should be corrected” (OSQ #150).

Without context, the LE communication may appear random, disordered, or incoherent. With context, the relevance of communication, including nonverbal engagement, was often perceived as goal-directed motor function that was atypical and contextually congruent: “My husband has been a musician all his life but 3 years ago could not remember how to play his drums. He, for the 20 years we have been married, has always played drums with his fingers on his legs, tabletops, etc. He quit doing that about 2 years ago. When 2 band members came to visit him after not seeing him for 6 months, he began to play the drums on his legs for a few minutes, leading me to believe he knew he played in a band with them and they triggered this response, although he did not speak to them or know them by name” (OSQ #33).

Guiding question 3: Do LEs include meta-awareness? Is insight about one’s cognitive impairment a distinct attribute of lucidity?

This code was defined as communication by the person living with ADRD during the LE that they were aware of their cognitive impairment and that they had insight about their health condition, diagnoses, or change in their health status. As volunteered by the caregiver, examples of this insight included the ability to communicate about their current state during the LE, and, in some cases, the cognitive clarity to discuss or react to their condition.

A caregiver who witnessed two different LEs for her spouse described them this way: “The 2017 experience was an enlightened period of time in which he realized for the first time that he had a problem and needed help. He had previously denied having any impairment. The 2020 experience was similar and opposite in that he had a moment of clarity and expressed anxiety about his progression. By this time, he had profound difficulty with coherent speech, but for a short while, we were able to talk about it” (OSQ #182). Another caregiver reported: “It was like the veil was lifted, he realized something was wrong with him and he was not going to survive it.” (OSQ #77). Similarly, another caregiver reported: “My husband asked me if he was dying and how soon he would die. He was lucid for the first time in 7 years—by which I mean we had a real talk about his present state. It was as if he was the same man 10 years prior to onset of dementia” (OSQ #3).

Guiding question 4: Are speech, language, and communication measures necessary and/or sufficient to identify the presence of an LE?

For this guiding question, we coded examples of different forms of communication (e.g., verbal vs. nonverbal) that caregivers reported witnessing in LEs. In addition, we also coded caregiver reports of communication quality (e.g., vocalizations or the pacing of speech) and coherence (e.g., logically organized and connected verbal and nonverbal language). These codes sometimes included caregivers’ interpretations of why these communication elements were important and indicated a cognitive shift in awareness.

Nonverbal communication was sometimes described as subtle and a connection or a sense of comprehension between the caregiver and the person living with ADRD: “He got less verbal, and his body sort of seized up… His hands kind of froze in a position…he couldn’t open his fingers, and he couldn’t smile. And then, he got so he couldn’t talk, but his eyes—you knew by his eyes. I would go in to feed him lunch [at the nursing home] and I went booting in there, always at high speed to feed him lunch. [Once] his son was there because it was his birthday month. And like he was excited. I mean he kept looking at [his son] and then he looked at me, and he did smile. And his expression was, ‘[My son] is really here,’ because [his son] did not come very often. Like that was a big response” (IR #7). Another caregiver used words such as “focused and certainty” to describe the change in non-verbal abilities that she associated with lucidity: “… the biggest difference I saw in my mother during a lucid moment was in her eyes. Mostly she had a nervous, unsure look in her eyes. During a lucid moment, her eyes were clear, a focused gaze and certainty in her expression” (OSQ #21).

Quality of the communication may also be an indicator of lucidity. One caregiver described it this way: “Mom came on the phone, her voice clear, strong and teasing and said, ‘You know, I don’t know what’s wrong with [the assistant]. She knows I have to have my coffee in the morning!’ Stunned to hear Mom so natural and relaxed, I went with it, and said, ‘I’m sure she’s just a little behind’… She hung up and shortly went back to her deteriorated condition” (OSQ #135).

From a caregiver’s perspective, quality of communication during an LE may also include a shift from non-communication and a seemingly disoriented state to “coherent communication,” in which another person is able to derive logic from the interaction. For example: “[He] had been diagnosed with early onset dementia. Now, [he] had not spoken for several years. I brought him [his favorite dish], the eggplant parmesan. He took one bite, looked at me, and said, [wife’s name]. Everybody in the place, like, dropped. They just stopped what they were doing, and said, ‘Did that really come out of him?’ It brought me to my knees. It brought me to tears… So the following week, I did the eggplant parmesan. I brought it in. I thought, ‘Oh, my gosh I’m gonna do it again.’ Nothing. Never happened again” (IR #1).

An example from another caregiver includes their interpretation of the logical organization of thought during the cognitive shift: “My mom called me by name and told me about how much she appreciated everything I had done for here. She then told me it was time to go home. She needed to go be with her mom and dad again. She asked me to take care of my dad and to not leave her alone after she passed” (OSQ #122).

Guiding question 5: Do LEs have intrinsic temporal markers or have a necessary magnitude of change in order to be recognized?

The ability to differentiate and discriminate core features of LEs from routine fluctuations in cognition has been proposed as a key to characterizing and defining LEs. In their data, wefound that some caregivers considered temporal dimensions of the LE (e.g., how long the LE lasted or how long it had been since the same level of clarity had been observable), and characterized the magnitude of change from a baseline to the LE as key characteristics of the LE.

Caregivers described various degrees of change during LEs, from small, quick, nuanced shifts in nonverbal alertness and visual connection to remarkable shifts in verbal language and abilities. Small changes often were described as a shift from glazed looks with no focus to relevant and intentional nonverbal communication that included smiling, touching hands, or making clear and intentional eye contact with the witness. For example, “I was visiting Mom…and had only 30 minutes to visit. She only became lucid as I was leaving. Her eyes opened wide, and she smiled and grabbed my hands. It was wonderful and heartbreaking at the same time” (OSQ #63). Similarly, “At my sister’s Christmas party in 2014, I handed my mom a cup of tea. She smiled, looked me in the eye, and very clearly said, ‘I love you.’ Her eyes were clear, there was no confusion or blankness in her face. That was the last time she ever verbalized those words to me. Within a few seconds, she glazed over again, and her face sagged with the usual dullness she’d had for about a year. The second time, I was feeding her dinner, and I told her I loved her. Again, she looked me right in the eye and nodded. Bright eyes, fully aware and full of meaning” (OSQ #17).

Some described short, but significant shifts from what was reported to be extreme cognitive impairment with difficulty communicating to clear and coherent interactions: “…she’s, like, laying there [outside during the summer] like-like a, you know, a cat just soakin’ up the sun with this big smirk on her face. And I just-I just kinda laughed, and I said, ‘Mom, do I have to pull that bikini out for you or what?’ And out of a dead sleep, she woke up, and she goes, ‘[Daughter], I don’t think anyone wants to see me in a bikini.’ And then she went back to sleep. And I sat on her bed, and I bawled for two hours straight. She hadn’t said my name in like three years” (IR #11).

Other unique conceptual considerations

Some descriptions by family members offered insight into additional conceptual challenges for differentiating LEs from routine cognitive fluctuations and other cognitive changes, such as symptoms associated with delirium (e.g., confusion, reduced awareness, and agitation). From the perspective of the caregiver, these LEs were considered unique because they were seen as lucid episodes embedded in periods of delirium or an episodic swing from a decline in awareness to lucidity. They reported changes that spanned from a baseline to a significant decline in awareness, sometimes with hallucinations, to the other extreme that included acute awareness and clarity, and then back to disordered, delirium-associated symptoms. For example, in one description, a caregiver described a parent’s recovered verbal abilities, which were logical and coherent, but also appeared to have end-of-life visions or hallucinations: “…it was the night before he passed. He had been in a morphine-induced like almost comatose state and not moving his arms or anything… all at once he just started talking and told my brother to go move the truck so that he could get the corn off the combine. And he turned around and looked like he was backin’ up the combine… he hadn’t moved his arms in forever, and then was throwin’ a fishing pole out ‘cause he loved to fish. And he was reelin’ it in-in. If we would ask him a question he would answer it… the thought of this man on morphine (because he had broken his back because he fell out of bed at the nursing home), you couldn’t tell that his back was broken because he was living his life through whatever dreams or hallucinations or lucidity or whatever you wanna call it was happening to him. He came back to life for four hours” (IR #2). In another report, amidst other fluctuating symptoms and cognitive changes, a caregiver described an embedded lucid episode: “My father had fallen at the nursing home. He had a mild brain bleed. My daughter and I visited him at the hospital and he was very agitated and restless. The nurse asked if I could help with his medication since he was refusing to take any meds. After some encouragement from me, he had a few moments of recognition. He took his meds and for the next few minutes was very responsive. My father held my hand and kissed it several times. He spoke to me with clarity and responded when I told him I loved him with a smile. A few minutes later he was back in his agitated state and not talkative” (OSQ #105).

Discussion

Establishing evidence-based conceptual and operational definitions is crucial for advancing an understanding of the epidemiology of LEs, including their possible catalysts or neurological pathways, and for providing guidance and support to caregivers and providers who care for those with ADRD (Eldadah et al., 2019; Peterson, Clapp, Largent, et al., 2022). Gilmore-Bykovskyi et al.’s PRF-L provides a foundational step toward identifying conceptual questions in LE definition and operational development (Gilmore-Bykovskyi et al., 2023). This study builds on that foundation by providing contextual insights from caregivers that illuminate potential LE features that may be definitionally essential or more variable. The findings of this study expound on the heterogeneity of LEs reported by caregivers (Gilmore-Bykovskyi et al., 2023; Ramirez et al., 2023; Teresi et al., 2023) and illustrate the challenges and complexity that researchers need to consider as they weigh optimal methods for capturing LEs and as they consider how to develop and test measures to label LEs with precision and coherence.

This research adds to this growing body of evidence and builds on previous reports (Batthyány & Greyson, 2021; Nahm & Greyson, 2009; Nahm et al., 2012) to show some common attributes across LEs that may help to refine quantitative measures to assess their occurrence and impact (Griffin et al., 2024; Ramirez et al., 2023). Caregivers perceived an unexpected shift in cognition, with some providing an interpretation of the magnitude of change from baseline and clarity in and context of emotion, behavior, or communication. Given the heterogeneity across caregiver reports, it is important to consider that moving forward, researchers collect data from larger samples and consider using quantitative assessments for a broad spectrum of indicators to help determine which indicators must be present to identify an LE, and which are more variable or have greater impact on the person living with ADRD and the episode’s observer/witness. In addition, because few studies have explored LEs by type of dementia, and some types, specifically dementia with Lewy bodies and Parkinson’s disease are characterized by fluctuating cognition (Ballard et al., 2001; Bradshaw et al., 2004), researchers should consider capturing cognitive fluctuations from people with Alzheimer’s disease and other types of dementia. These data could be used to develop robust methods and standards to distinguish LEs from expected or routine cognitive fluctuations.

Similar to previous findings (Gilmore-Bykovskyi et al., 2023; Karlawish et al., 2024; Ramirez et al., 2023; Teresi et al., 2023), we found that reports of LEs were characterized by a heterogenous set of goal-directed behaviors with both verbal and nonverbal communication, suggesting this is a necessary indicator for LEs. Some LEs included recognition or awareness of people, emotions, surroundings, and their own limitations, suggesting that these may be indicators of LE, but perhaps not necessary to confirm an episode as lucid. Also consistent with other studies (Benson et al., 2023; Teresi et al., 2023), we found that LEs occurred across different levels of cognitive impairment, as reported by the caregiver, with varying levels of complexity in communication or behavior, and caregivers often reported LEs with subtle or nuanced changes or changes that required personal context to attribute it as a shift toward lucidity.

The findings of this study provide some guidance on developing definitions and measurement of LEs. Based on findings from Guiding Question 1, for example, where the researchers found descriptions of LEs across caregiver reported levels of cognitive impairment, researchers may want to consider including caregivers across different stages of ADRD in order to capture episodes earlier in the trajectory, not just at the end of life. Similar to others (Benson et al., 2023), the findings include reports of people in seemingly later stages of dementia who experience LEs where they profess their love for their family and those with reportedly moderate cognitive difficulty whose short-term recovered abilities included expressions of more complex emotions than the caregivers had thought was retrievable, including distress and the unpleasantness of their lived experience.

Results from Guiding Question 2 similarly demonstrated heterogeneity of recovered abilities during LEs and often a need for interpretation of the significance of the event or demonstration of personality congruence. With the underlying pathologies and clinical presentation of ADRD so diverse (Warren et al., 2012), it is not especially surprising that recovered abilities are diverse. The researchers found, however, examples where only a family member or someone with deep knowledge of the personality or personal history of preexisting and current abilities would be able to recognize a recovered ability as meaningful. This further underscores the inherently observer-reliant nature of LE detection, as people with advanced cognitive impairment lack the ability to self-report their experiences. Rather than being universally defined events, LEs are identified through noticeable shifts in behavior and communication that are meaningful within a specific personal and relational context. As with other observer-reliant neurological phenomena (such as delirium), caregiver and clinician accounts are the primary sources of data, despite potential recall bias. The data in this study support Gilmore-Bykovskyi’s conjecture that, at the current stage of conceptual development of LE definitions and measurement, idiopathic approaches (particular to the individual) may be preferable over nomothetic (descriptive of a group of individuals) for characterizing LEs (Gilmore-Bykovskyi et al., 2023). As the strength of evidence builds, there will likely be a need for complementary nomothetic approaches in order to understand the epidemiology of LEs in ADRD.

Relevant to Guiding Question 3, the researchers found some caregivers reported elements of meta-awareness during the LE. In the researchers’ reports, they found that the people living with ADRD not only stated their awareness of their condition but that there was sometimes a discussion with the witness about it. Assessing whether meta-awareness occurs (yes or no) may help determine if LEs with meta-awareness are a distinct type of LE that differs from LEs without (Griffin et al., 2024; Morris & Bulman, 2020) or associated with a specific type or clinically determined stage of dementia. To better understand meta-awareness as an attribute of LEs, researchers will need to be rigorous and thorough about meta-awareness measurement, especially if LEs are reported by witnesses and not recorded. This will help to reduce bias by distinguishing meta-awareness interpreted by witnesses (e.g., “It seemed like he was saying to me that he understood his current physical state) from more explicit experiences (e.g., “He said this to me: I have not been able to talk to you, but now I can.”).

Relevant to Guiding Question 4, caregivers in this study offered examples of LEs with recovery of verbal and nonverbal abilities, with different interpretations of the communication and highlight the challenges of determining and distinguishing a LE from a non-LE. The findings of this study illuminate whether elements of communication, such as coherence and logic, are necessary for defining an LE, or if recovery of any lost ability, even if it is not coherent or logical, should be included with other indicators in a different type of lucidity. One of the examples showed recovery of nonverbal abilities without personal interaction (e.g., driving a tractor, fishing), leaving questions about whether this episode could be considered an LE (Peterson, Clapp, Harkins, et al., 2022; Peterson, Clapp, Largent, et al., 2022), a “deathbed vision” at the end-of-life (Kellehear, 2017), hallucinations associated with delirium or a behavioral symptom of ADRD, or all three. It is also possible that some forms of nonverbal communication are more aligned with implicit memory than explicit memory (Sabat & Warren, 2023), highlighting the need for research to measure and distinguish lucidity from other forms of cognitive fluctuation or routine activities.

The findings of this study also suggest that not just the nature of communication (verbal or non-verbal) but the interpretation of that communication by the witness is important. Interpretation, however, may introduce bias that researchers need to consider. Witnesses’ desires to connect and communicate with the person living with ADRD may lead them to identify subtle shifts in cognition as an LE even if those shifts are within a pattern of routine fluctuations. For researchers, robust methods that capture quantitative and qualitative details on LEs and on those that witness them will allow for more precise methods to adjudicate the presence of an LE. Capturing LEs using different methods than caregiver or healthcare provider reports, such as video recordings (Gilmore-Bykovskyi et al., 2024) may be helpful in reducing the potential for these witnesses to conflate a LE from a non-LE but may still require a contextual understanding to adjudicate.

Relevant to Guiding Question 5, the findings provide insight into some distinct out-of-the-ordinary changes in cognition that were noticeable and remarkable to caregivers, and their variation from routine and expected fluctuations was a characteristic they emphasized. One limitation, however, is that we did not have corresponding data to compare the magnitude of change captured quantitatively (e.g., how long the events were or how much they changed from baseline cognition) with qualitative reports. Having both types of data in future research will provide a clearer understanding of whether a certain magnitude of change is essential for defining LEs. Moreover, caregiver interpretation of cognition is rather subjective and introduces some noise into measuring the magnitude of cognitive change within a LE. Having a more precise measure of cognition is a priority for future research on LEs.

One novel finding is that cognitive fluctuations may include changes from an acute state with disordered communication and delirium-related symptoms to awareness with clarity. Some reports showed elements of delirium preceding and/or following an LE, possibly suggesting that, just as delirium and dementia co-exist (Fong et al., 2015), so may delirium, dementia, and lucid episodes. Research will be needed to modify measures that assess a spectrum of cognitive fluctuations, not just those indicating lucidity, which could include testing the use of a sliding scale from disordered and confused to awareness and clarity and the magnitude of change on that continuum.

Limitations

This study also has important limitations to consider. First, without an explicit or evidence-based definition of LEs, it is possible that some episodes described could eventually not be considered LEs, and because of the reliance on caregiver reporting for dementia type and cognitive severity, that some reported LEs were more consistent with routine cognitive fluctuations in specific types of dementia (e.g., Lewy body dementia) or with what would clinically be determined as earlier stages of ADRD that have more variable cognitive fluctuations. However, at this early stage in the conceptualization of the phenomenon and in response to critiques to not impose overly rigid definitional criteria (Peterson, Clapp, Largent, et al., 2022), the researchers have cast a wide net of what is considered an LE and assumed that caregivers who report LEs, indeed, witnessed an LE. Second, it is also possible that at this early stage of inquiry, additional conceptual questions are still needed to advance the operationalization of LEs. Third, it is possible that a structured interview that included specific questions about antecedents, consequences, and LE attributes would result in findings different from those of the semi-structured approach the researchers used.

Conclusions

The results of the qualitative interviews provide additional insights into caregivers’ experience of what they identified as lucid episodes among those for whom they provided care. The existing guiding questions are helpful in further describing the clinical phenomenology of lucid episodes, and the data provide examples of how these questions can be used to advance methods to study LEs and measures to assess LEs and their impact. Rich, detailed accounts from caregivers illustrate the complexity of these experiences in ways that quantitative studies may miss. The results of this study add to the understanding of this phenomenon and suggest the need to refine how the researchers study it in order to better capture the complex and varied clinical expression of lucid episodes as experienced by caregivers of people living with ADRD.

Supplementary material

Supplementary data are available at The Gerontologist online.

Funding

This study was supported by funding from the National Institute on Aging (R21AG069767 and R33AG069767, Definition and Caregiver Appraisal of Paradoxical Lucidity in Dementia; PI: J.M.G.). K.K. was supported by the New Faculty Startup Fund from Seoul National University. Y.L. was supported by a mentored career development award from the National Institute on Aging (1K01AG081566, Cognitive Health and Modifiable Factors of Daily Sleep and Activities among Dementia Family Caregivers). Funders did not influence the development or presentation of this study despite authors’ affiliation with the funder.

Conflict of interest

None declared.

Data availability

Interview guide is available in the supplementary materials. Qualitative data is not available. This study was not pre-registered.