key: cord-0763568-0exa85zr
authors: Gorgoni, Maurizio; Scarpelli, Serena; Alfonsi, Valentina; De Gennaro, Luigi
title: Dreaming During the COVID-19 Pandemic: A Narrative Review
date: 2022-05-25
journal: Neurosci Biobehav Rev
DOI: 10.1016/j.neubiorev.2022.104710
sha: 7aff4af381e7b11a4adb6a17ddb9f8a7dd4157e8
doc_id: 763568
cord_uid: 0exa85zr

According to several influential models, dreams can be affected by state- and trait-like factors, sleep features, and diurnal experiences. The COVID-19 pandemic globally affected daily habits, emotional experiences, and sleep. Previous studies suggested an influence of collective traumatic events on dreaming. Starting from these premises, several studies assessed the effect of the pandemic on dreams. This paper aims to review findings concerning the oneiric activity during the COVID-19 pandemic. We report pandemic-related changes in dreams and nightmares, and we consider the possible factors associated with dreaming. Moreover, we provide results about changes in the oneiric activity in different phases of the pandemic. The reviewed findings suggest a pandemic-related enhancement of dream and nightmare frequency, emotional intensity, and distressing contents, modulated by modifications in restrictive measures and associated with diurnal experiences, emotional status, and sleep pattern. We highlight several methodological issues and a large heterogeneity in the present literature, limiting results’ generalizability. However, we provide possible interpretations of the most consistent findings in light of the main theoretical frameworks about dreaming.

contents/themes analysis of specific dreams recorded (e.g., written, audio-recorded) by the participants, and dream items/questionnaires/diaries (24 studies), which allowed to score dream frequency, qualitative features, and contents. Also, 16 studies assessed sleep features, sleep disorders, or sleep-related variables, and 21 papers report on psychological/emotional/clinical variables. Twenty-three studies were performed on the general population, while other studies were focused on specific samples: university students (3 studies), undergraduates/postgraduates (1 study), secondary school students (1 study), adolescents (1 study), healthcare workers (1 study), narcoleptic patients (1 study), COVID-19 patients (1 study). Table 1 about here

Changes in oneiric phenomenology during the pandemic compared to a non-pandemic condition were assessed with different methods: direct questions about the perception of pandemic-related changes in dreams (Schredl and Bulkeley, 2020; Guerrero-Gomez et al., 2021) and nightmares (Pesonen et al., 2020; Guerrero-Gomez et al., 2021) , retrospective scoring of dream features (Conte et al., 2022; Gorgoni et al., 2021a; Frankl et al., 2021; Solomonova et al., 2021; Scarpelli et al., 2022b) and nightmare features (Musse et al., 2020; Solomonova et al., 2021; Scarpelli et al., 2022b) before and during the pandemic, comparison with a population-based pre-pandemic sample (Scarpelli et al., 2021a; Kilius et al., 2021) , and comparison with a non-pandemic control group (McKay and DeCicco, 2020; Mota et al., 2020; Wang et al., 2021) . Changes in dream frequency during the pandemic in the general population were reported by several studies (Schredl and Bulkeley, 2020;  J o u r n a l P r e -p r o o f Conte et al., 2022; Gorgoni et al., 2021a; Frankl et al., 2021; Scarpelli et al., 2021a; Solomonova et al., 2021; Scarpelli et al., 2022b) . Conte and co-workers (2022) found that approximately half of the participants reported increased or decreased dream frequency during total and partial lockdown compared to pre-lockdown periods. The most frequently reported result is represented by increased dream frequency during total lockdown compared to the pre-pandemic period (Schredl and Bulkeley, 2020; Conte et al., 2022; Gorgoni et al., 2021a; Frankl et al., 2021; Scarpelli et al., 2021a; 2022b; Solomonova et al., 2021) . This result was also confirmed in secondary school students (Guerrero-Gomez et al., 2021) and COVID-19 patients (Scarpelli et al., 2022b) . Changes in nightmare frequency during the pandemic, mainly in the direction of a pandemic-related increase, were reported by 3 studies in the general population (Musse et al., 2020; Frankl, 2021; Scarpelli et al., 2021a) , one study in secondary school students (Guerrero-Gomez et al., 2021) and COVID-19 patients (Scarpelli et al. 2022b ).

Pandemic-related modifications in contents and qualitative features of the oneiric activity have been observed by 5 studies in the general population (Mota et al., 2020; Schredl and Bulkeley, 2020; Conte et al., 2022; Gorgoni et al., 2021a; Wang et al., 2021) and two papers on small groups of university students (McKay and DeCicco, 2020; Kilius et al., 2021) .

Specifically, several studies found changes in the emotional experience of dreams, in the direction of greater emotional intensity (Gorgoni et al., 2021) , more negative tone, or more frequent negative emotions in dreams (Schredl and Bulkeley, 2020; Conte et al., 2022; Gorgoni et al., 2021a) compared to pre-lockdown period. Conte and co-workers (2022) also observed that approximately half of the participants reported increased or decreased dream length and vividness during total and partial lockdown compared to pre-lockdown J o u r n a l P r e -p r o o f periods, and one study found a pandemic-related increase in dream vividness, bizarreness, and length compared to the pre-pandemic period (Gorgoni et al., 2021a) .

The analysis of changes in contents of dreams collected during the pandemic points to a) more animal imagery, location changes, virus-related content, food imagery, and head dream content in university students (McKay and De Cicco, 2020) ; b) a greater proportion of anger-and sadness-related words and higher average semantic similarities to the term "contamination" and "cleanness", associated with mental suffering linked to social isolation (Mota et al., 2020) ; c) aggressive interactions in female university student (Kilius et al., 2021) ; d) increased frequency of threatening (non-aggressive) events (Wang et al., 2021) .

Several variables were associated with specific pandemic-related changes in the oneiric activity. Female gender was related to a pandemic increase in dream frequency (Gorgoni et al., 2021a , Guerrero-Gomez et al., 2021 , emotional load and bizarreness (Gorgoni et al., 2021a) , negative dream tone, and pandemic-related dreams (Schredl and Bulkeley, 2020), nightmare frequency (Pesonen et al., 2020 , Guerrero-Gomez et al., 2021 Kilius et al., 2021) , and physical aggression in dream contents (Kilius et al., 2021) . Increased stress was associated with boosted nightmare frequency (Pesonen et al., 2020) . People most strongly affected by the pandemic exhibited heightened dream recall, more negative dream tone, and pandemic-related dreams (Schredl and Bulkeley, 2020). Younger age was associated with increased dream recall (Schredl and Bulkeley, 2020), but older participants had greater nightmare frequency in a small group of university students (Guerrero-Gomez et al., 2021) . Higher education was associated with increased dream recall and COVID-19related dreams in one paper (Schredl and Bulkeley, 2020) . In a single study (Gorgoni et al., 2021) it was observed that a) participants with poorer sleep quality had a greater increase J o u r n a l P r e -p r o o f in dream frequency and emotional load; b) individuals with PTSD-related nocturnal sleep behaviours had a greater increase in dream frequency, emotional load, vividness, and bizarreness; c) participants with depressive symptoms had a greater increase in dream frequency, emotional load, and vividness. Another study found that the proportion of participants reporting poor sleep quality, nightmares, and frequent sleep problems was higher during the pandemic in both low and high dream recall frequency groups, but such increase was greater for participants with high dream recall frequency (Frankl et al., 2021) .

Finally, in secondary school students, one study found that different indexes of pandemicrelated subjective distress and emotional reactions were positively associated with nightmare increase, Italian participants more likely reported more nightmares compared to Croatian and Romanian, and experiencing discomfort/sadness was related to dream increase (Guerrero-Gomez et al., 2021) .

One study assessed the predictors of bidirectional changes (i.e., increase and decrease) in dream variables during total and partial lockdown compared to pre-lockdown periods (Conte et al., 2022) . During the total lockdown, a) younger age predicted decreased and increased dream frequency, length, and vividness; b) female gender predicted decreased and increased dream frequency and vividness; c) worsened sleep quality predicted decreased dream frequency, length, and vividness; d) increased dream negative emotionality predicted increased dream frequency, length, and both increased and decreased vividness; e) delayed sleep midpoint, negative mood, and moderate stress respectively predicted increased dream length, decreased and increased dream vividness.

During the partial lockdown, a) younger age predicted decreased dream frequency and length, b) positive mood predicted decreased dream frequency and was negatively associated with increased dream vividness, c) increased negative dream emotionality J o u r n a l P r e -p r o o f predicted decreased and increased dream frequency and length. The increase in negative emotionality was predicted by female gender, higher stress, more negative mood and worsened sleep quality in total lockdown, and worsened sleep quality during the partial lockdown. The presence of COVID-19 related dreams was predicted by female gender, higher general fear, increased dream vividness, and increased dream negative affect during total lockdown, and higher general fear during partial lockdown (Conte et al., 2022) .

Several studies found higher dream frequency (Iorio et al., 2020; Gorgoni et al., 2021a; Frankl et al., 2021; Scarpelli et al., 2021a) or a greater percentage of high dream recallers (Sommantico et al., 2021) in women than men during the pandemic. Younger age was related to greater dream frequency (Gorgoni et al., 2021a; Frankl et al., 2021; Scarpelli et al., 2021a ) and a greater percentage of high dream recallers (Sommantico et al., 2021) .

Considering the emotional status, one study found an association between PTSD symptoms and higher dream recall frequency (Frankl et al., 2021) , and participants with moderate-to-severe stress levels showed greater dream frequency in another study (Solomonova et al., 2021) . The presence of depressive symptoms was associated with higher dream frequency in one paper (Gorgoni et al., 2021a) . Frankl and co-workers (2021) also found a positive relationship between high dream recall frequency and depression, but when PTSD symptoms were included in the regression model, the direction of the relationship between dream recall frequency and depression turned negative, and the same phenomenon was found for anxiety. Two studies found that several environmental J o u r n a l P r e -p r o o f factors were associated with high dream frequency: territorial area (i.e., north of Italy), living alone during the lockdown (Gorgoni et al., 2021a) , and not having children (Scarpelli et al., 2021a) . Media exposure was found positively associated with threatening dream frequency, with anxiety as a partial mediator (Guo and Shen, 2021) . Finally, three studies found an association between great dream frequency and sleep variables: higher sleep duration (Gorgoni et al., 2021a; Scarpelli et al., 2021a) , higher sleep disturbance (Gorgoni et al., 2021a) , nightmares, sleep talking, sleep maintenance problems, and RBD (Frankl et al., 2021) . Greater dream recall was also associated with sleepiness in a group of narcoleptic patients (Scarpelli et al., 2021b) .

The emotional intensity of dreams during the pandemic was greater in females (Iorio et al., 2020; Gorgoni et al., 2021a; Sommantico et al., 2021) , which also reported more negative emotional tone or more frequent negative emotions in dreams (Barrett, 2020; Iorio et al., 2021; Gorgoni et al., 2021a) . Dream emotional intensity was also associated with younger age, depressive symptoms, territorial area (i.e., north of Italy), living alone during the lockdown, poorer sleep quality, greater sleep duration, sleep disturbance, and daytime dysfunctions (Gorgoni et al., 2021a) , job change and asking for help from mental health professional (Scarpelli et al., 2021c) . Moreover, in one study greater emotional intensity was observed in people who stopped working, reported modification of the sleep onset, and changed habits at morning awakening (Scarpelli et al., 2021a) . One paper reported more frequent dream negative emotions in younger adults, and participants with poor sleep quality, nocturnal disruptive behaviors, anxiety, and depressive symptoms (Gorgoni et al., 2021a) , and a more negative emotional tone was observed in people who J o u r n a l P r e -p r o o f had stopped working, reported modification of the sleep onset and changed habits at morning awakening (Scarpelli et al., 2021a) , and in adults compared with adolescents (Sommantico et al., 2021) . The negative emotional tone was also associated with forced quarantine, having relatives/friends infected by COVID-19, and asking for help from mental health professional (Scarpelli et al., 2021c Several studies used different techniques to qualitative or quantitative analyse dream contents, themes, and narrative styles during the pandemic. Barrett (2020) collected any J o u r n a l P r e -p r o o f dreams related to COVID-19, assessing sex differences in emotions and body concerns from the Linguistic Inquiry and Word Count (LIWC). The author found that women showed lower positive emotions in dreams and greater rates of negative emotions, anxiety, sadness, anger, body content, references to biological processes, health, and death.

Men exhibited a higher score for the LIWC variable, while positive emotions, negative emotions, anxiety, and death had a similar trend to the one observed in women, but with lower significant levels (Barrett, 2020) . Dream related to COVID-19 has been also compared to a large number of tweets (considered an expression of waking experiences) mentioning the pandemic, using a deep-learning algorithm for the extraction of mentions of medical conditions (Šćepanović et al., 2022) . Results suggest that a) health expression common to dream dataset and tweets referred to typical COVID-19 symptoms, while those that distinguished dreams and tweets reflected differences in thought processes: tweets reflected linear and logical thoughts, describing realistic symptoms and related disorders; dreams reflected conditions unrelated to the virus and conditions of surreal nature.

Iorio and co-workers (2020) collected the most recent dream, assessing the common presence of qualitative dream aspects and analysing their content using the Grounded Theory Model (Glaser & Strauss, 1967) , a method that allows the development of a theoretical account concerning the general features of a topic while grounding the account in empirically observation or data (Martin and Turner, 1986) . The main findings show that a) women exhibited greater dream emotional intensity, predominantly negative dream emotional tone, greater negative emotions and sensory impressions than men, b) 20% of the reported dreams had explicit COVID-19 references, c) participants knowing people affected by or died for COVID-19 exhibited greater emotional intensity and sensory J o u r n a l P r e -p r o o f impressions in dreams, d) the most recent dreams were more likely set in external locations and showed negative emotions (mainly for dangerous, violent, and frustrating situations). The same research group collected the most recent dream around the end of the lockdown in adults and adolescents, also performing a stemming process and a Thematic Analysis of Elementary Contexts (Sommantico et al., 2021) . The authors found that females reported longer dreams, lower positive emotions, higher negative emotions, and greater presence of sensory impressions than males. Adult participants reported longer dreams, higher negative emotions, and higher presence of sensory impressions than adolescents. Clusters of adolescents' more recent dream included home confinement, school, friends and boyfriends, and death of family members. Clusters of adults' more recent dream included home confinement, nostalgia for happiness, bodily sensations, dream evaluation, and compliance with health and safety norms.

Pesonen and collaborators (2020) assessed the content of dream collected during the lockdown using network and cluster analysis. Results revealed 33 dream clusters, including 20 bad dream clusters of which 55% were pandemic-specific. The dreamassociation networks were more accentuated for those who reported an increase in perceived stress.

Guerrero-Gomez and collaborators (2021) asked secondary school students to report "an extraordinary dream", in which the authors identified pandemic contents. A direct reference to COVID-19 related topics was found in 14.2% of reported dreams. Younger age, female gender, and worries about another lockdown were associated with the report of a pandemic-related dream.

In the study of Borghi and co-workers (2021), the authors performed a thematic analysis, a qualitative coding process that allows the identification, analysis, and aggregation of Two studies from Mariani and co-workers (2021a; 2021b) also focused on dream themes.

In the first study (Mariani et al., 2021a) , the authors collected dreams from an online blog during the lockdown and analysed dream themes using different computerized referential processes linguistic measures, revealing three dream clusters: symbolizing process, arousal of emotional activation, reflection/reorganizing elaboration. In the second study (Mariani et al., 2021b) , they applied computerized referential process linguistic measures and Affect Salience Index, a measure of arousal in written texts (Fornari, 1976; Salvatore and Freda, 2011; Valsiner, 2021) , on dreams and waking thoughts collected during the lockdown. The authors found that a) Affect Salience was present in dreams and waking J o u r n a l P r e -p r o o f thoughts, b) Referential Activity was higher in dreams, while Reflection and Affect words were higher in waking thoughts, and c) a greater symbolization process was observed during dreams and higher emotional distance in waking thoughts.

Solomonova and co-workers (2021) assessed dream topics and themes during the lockdown using a modified version of the Typical Dream Questionnaire (Nielsen et al., 2003) that also includes 4 pandemic-related concerns. The results showed that the most common dream themes during the pandemic were centered around the topics of inefficacy, human threat, death, and pandemic imagery. Higher stress, anxiety, and depression levels were associated with the frequency of dreams contents about the pandemic, inefficacy, and death.

Nightmare frequency during the pandemic was associated with (and was more frequent 

Nightmare frequency was also assessed in specific populations. One study on healthcare workers found that individuals with frequent nightmares during the pandemic reported higher worry about being infected and family's infection of COVID-19, lower confidence in being cured if infected and sense of competence, a greater percentage of poor mental health, higher General Health Questionnaire score, reduced sleep duration, higher diurnal sleepiness, and poorer sleep quality (Lin et al., 2021) . Moreover, reduced sleep duration and reduced habitual sleep efficiency were associated with frequent nightmares (Lin et al., 2021) . In a small group of narcoleptic patients, nightmare frequency was related to the female gender, longer sleep duration, higher intra-sleep wakefulness, and sleepiness (Scarpelli et al., 2021b) . In another study, individuals with COVID-19, compared to a control group, exhibited higher nightmare frequency, anxiety, depression, PTSD scores, and lower quality of life, quality of health, and wellbeing (Scarpelli et al., 2022b) . In the COVID-19 group, greater nightmare frequency was predicted by higher dream recall frequency, higher anxiety, insomnia symptoms, high PTSD risk, lower sleep duration, and younger age (Scarpelli et al., 2022b) .

Concerning qualitative aspects, in one study nightmare distress was greater in individuals who stopped working during the pandemic, having relatives or friends infected by or died for COVID-19, and exhibiting modification of sleep onset, morning awakening habits, and napping habits (Scarpelli et al., 2021a).

Nightmare contents have been also assessed. Previous psychiatric conditions, use of sleep medication, and younger age were found associated with nightmares characterized by pandemic contents (Musse et al., 2020) . Kennedy and co-workers (2022) found that different nightmare contents were selectively associated with greater general COVIDrelated stress, worsened sleep, and worsened middle-of-the-night insomnia, while J o u r n a l P r e -p r o o f depression and anxiety were positively associated with all of the investigated nightmare themes. Finally, Alghamdi and co-workers (2022) observed in university students that the most common nightmare theme was not related to COVID-19. Two studies assessed the oneiric activity during the total and partial lockdown associated with the first and second waves of the pandemic. Conte and co-workers (2022) showed that the partial lockdown was characterized by a reduced proportion of participants reporting increased or decreased dream frequency, increased dream length and vividness, without a difference in the proportion of participants reporting COVID-19 related dreams, compared with the total lockdown period. Scarpelli and co-workers (2021c) observed that, compared to the first pandemic wave, participants had reduced dream recall frequency, nightmare frequency, lucid dream frequency, emotional intensity, nightmare distress, and greater negativity of the dream emotional valence in the second wave. Specific differences in post-traumatic growth, nocturnal disruptive behaviors, sleep, and sleep-related 

Considering the relatively small period of interest, a large number of studies focused on the oneiric activity during the pandemic, denoting the strong interest in this topic. The larger part of the considered studies were cross-sectional designs aimed at the characterization of dreams and their association with diurnal habits, emotional and clinical conditions, and sleep during the pandemic, or the retrospective assessment of changes in the oneiric activity compared to pre-pandemic/non-pandemic conditions. Only a few studies assessed longitudinal modifications in dreaming during the pandemic.

Dream items/questionnaires/diaries were more frequently used than dream reports.

Many studies were performed in the general population, while the assessment of dreams and nightmares in specific samples (i.e., healthcare workers, narcoleptic patients, and COVID-19 patients) was rarer.

Nightmares were widely assessed, but it should be considered that the reviewed literature is characterized by a lack of clarity and uniformity about the operational definitions of to the participants, and only a few studies distinguished "nightmares" and "bad dreams".

In this view, many results about pandemic nightmares likely encompass both nightmares and bad dreams. Concerning lucid dreams, only 4 studies assess them during the pandemic using different experimental designs, and one of them was focused on narcoleptic patients. Therefore, it is difficult to determine the effect of the pandemic on lucid dreaming at present.

From a methodological standpoint, the reviewed studies are characterized by several issues, which lead us to recommend great caution in the interpretation of these findings.

Overall, the present literature is characterized by a large heterogeneity concerning a) the methods and instruments used to collect and analyse the oneiric activity, b) the oneiric features considered, particularly concerning dream contents and themes, c) the clinical, sociodemographic, environmental, and sleep measures collected as possible predictors of dream features, d) the periods of the pandemic considered for data collection, e) the objectives of the study, f) the epistemological framework for the assessment of dream contents, themes, and narrative styles, g) the composition of the sample, h) the statistical models. As a consequence, studies are hardly comparable, and many findings have been not replicated (or they have been partially replicated using different methodologies).

The reported studies have been conducted during the pandemic, mainly in periods of forced isolation, leading to wide use of online recruitment and data collection strategies. This method can introduce a self-selection issue, attracting a large number of individuals with mental health problems, sleep alterations, or greater interest in dreams. Moreover, J o u r n a l P r e -p r o o f many studies were characterized by an unbalanced sample concerning several variables.

The sample size of the reported studies is extremely variable (min: two groups of 19 participants; max: 19355 participants), and it is worth noting that several studies were conducted on very small samples. These issues limit the generalizability of the results.

The available cross-sectional retrospective data mainly support the notion of an increased dream and nightmare frequency during the pandemic compared to pre-pandemic/non- 

The influence of sociodemographic and pandemic-related variables has been observed in several studies. The effect of gender on the oneiric activity has been frequently reported using different methods, mainly in the direction of greater dream frequency, emotional (negative) intensity, length, vividness, and pandemic-related contents (Barrett et al., 2020; Iorio et al., 2020; Gorgoni et al., 2021a; Frankl et al., 2021; Scarpelli et al., 2021a; Sommantico et al., 2021; Guerrero-Gomez et al., 2021) , and nightmare frequency (Musse et al., 2020; Scarpelli et al., 2021a; Alghamdi et al., 2022) . The association between nightmare J o u r n a l P r e -p r o o f frequency and female gender has been also observed in narcoleptic patients (Scarpelli et al., 2021b) . The effect of age has been frequently reported. Specifically, younger age has been associated with greater dream frequency, qualitative dream features, and pandemicrelated contents (Gorgoni et al., 2021a; Frankl et al., 2021; Scarpelli et al., 2021a; Sommantico et al., 2021; Guerrero-Gomez et al., 2021) Clearly, the above-mentioned unbalance in the sample of many studies, with an overrepresentation of several sociodemographic categories and a possible self-selection bias, represents e relevant methodological problem while assessing the relationships between these variables and dreaming. Crucially, in the vast majority of the studies, the sample included a large prevalence of females. In this view, it could be argued that the strong influence of gender on pandemic dreams and nightmares may represent a by-product of the large prevalence of females in the reviewed studies. Nevertheless, the existence of sex differences in dreaming activity has been widely observed also in pre-pandemic studies J o u r n a l P r e -p r o o f (Schredl and Reinhard, 2008; Schredl, 2010) . In particular, Schredl and Reinhard (2008) found in their meta-analysis a robust gender difference in dream recall only when collected outside of the sleep laboratories (i.e., without any control of the preceding sleep stage). 

The absence of information on pre-pandemic sleep and mental health features in the larger part of the studies should be highlighted since it strongly limits the possibility to control for the influence of pre-existing sleep and psychopathological problems on dreaming during the pandemic and reliably assess the relationship between these variables and the oneiric activity. Specifically concerning sleep measures, it should be considered that the reviewed studies lack of objective sleep assessment (i.e., polysomnographic, actigraphic), limiting an accurate evaluation of the relationship between sleep and dreaming during the pandemic. Furthermore, the possible relationship between different dream features has been scarcely considered in studies conducted during the pandemic. According to the salience hypothesis (Cohen and McNeilage, 1974) , the subjective impact of dreams may strongly affect dream recall frequency. However, from an opposite standpoint, it could be hypothesized that a larger number of recalled dreams may lead to a subjective perception of increased emotional intensity and qualitative properties. Therefore, results on parallel quantitative and qualitative dream changes during the pandemic (e.g., Gorgoni et al., 2021a; 2022 ) may be at least in part influenced by the relationship between dream features and may be interpreted in two ways: a) the greater emotional intensity of dreams, likely related to modifications in diurnal emotional experiences, led to a higher dream recall frequency; b) a greater dream production and/or recall, associated with diurnal experiences and/or alterations of the sleep pattern, led to a subjective perception of greater dream emotional intensity. Further studies are needed to disentangle this point.

The most consistent longitudinal finding is represented by reduced dream and lucid dream frequency during post-lockdown periods characterized by eased restrictive Considering the time passed during a lockdown period, Alfonsi and co-workers (2022) found an initial increase followed by a stabilization of dream recall frequency during the first Italian lockdown, and Goncalves and collaborators (2022) 

Overall, beyond the reported methodological issues characterizing the reviewed studies, The observed dream changes compared to pre-pandemic periods, the relationship with several facets of the diurnal experience, and the modification of the oneiric activity in different phases of the pandemic can be interpreted in the framework of the "continuity J o u r n a l P r e -p r o o f hypothesis", which claims that dreams reflect waking experiences, mental activity, and emotions (Schredl, 2006; Domhoff, 1996) , pointing to the existence of a continuum between waking and sleeping mental/emotional/neurobiological processes (Scarpelli et al., 2019a; 2021d) . Several findings support the notion that dream contents may mirror different As previously observed, while several studies revealed explicit pandemic-related contents/themes, the assessment of their prevalence led to conflicting findings (McKay and DeCicco, 2020; Musse et al., 2020; Mota et al., 2020; Iorio et al., 2020; Sommantico et al., J o u r n a l P r e -p r o o f 2021; Pesonen et al., 2020; Borghi et al., 2021; Kennedy et al., 2022; Solomonova et al., 2021; Guerrero-Gomez et al., 2021; Alghamdi et al., 2022) . Beyond methodological differences between these studies, it is worth noting that dreams do not necessarily incorporate the explicit episodic memories, more likely extracting the gist of the salient experiences (Eichenlaub et al., 2017; Malinowski and Horton, 2014; Wamsley and Stickgold, 2011) in an associative way. The common observation of pandemic-related increase or high frequency during the pandemic of distressing, emotionally negative contents in the oneiric activity, not necessarily containing an explicit pandemic reference, is in line with this view (Mota et al., 2020; Musse et al., 2020; Iorio et al., 2020; Kennedy et al., 2022; Kilius net al., 2021; Sommantico et al., 2021; Pesonen et al., 2020; Solomonova et al., 2021; Wang et al., 2021; Alghamdi et al., 2022) .

The reviewed literature also points to a role of sleep features on dream and nightmare recall during the pandemic. As observed, several indexes of worse sleep quality appear associated with greater dream/nightmare frequency and emotional intensity, specific dream qualitative features and contents and greater pandemic-related changes in the oneiric production (Gorgoni et al., 2021; Frankl et al., 2021; Kennedy et al., 2022; Scarpelli et al., 2021a; 2021b; 2022b; Lin et al., 2021; Alghamdi et al., 2022) . These findings are consistent with the hypothesis that a greater level of arousal during sleep and intra-sleep wakefulness facilitate the encoding of the oneiric experience, more likely leading to a successful dream recall (De Gennaro et al., 2010; Koulack, Goodenough, 1976) . Indeed, both healthy and clinical samples exhibit an association between greater dream recall and more fragmented sleep (van Wyk et al., 2019; Polini et al., 2017; Schredl, 2009 ).

Furthermore, several findings suggest that a greater electrophysiological desynchronization promotes dream recall (Scarpelli et al., 2017; 2020; Siclari et al., 2017;  J o u r n a l P r e -p r o o f 2018). The pandemic had a strong and complex impact on sleep (Partinen et al., 2021; Cellini et al., 2020; Franceschini et al., 2020; Alfonsi et al., 2021; Morin et al., 2021; Gorgoni et al., 2021b; 2021c; Salfi et al., 2021) , more frequently in the direction of a worse sleep quality (Jahrami et al., 2021; 2022) . Therefore, it is possible that the increased frequency and intensity of the oneiric activity during the pandemic may be at least in part a consequence of greater arousal during sleep associated with the frequent reduction of sleep quality, which would facilitate the process of dream encoding and recall. It should be also considered that many studies point to changes in the sleep timing during the pandemic, mainly represented by delayed bed time and rise time (Cellini et al., 2020; Gao and Scullin, 2020; Marelli et al., 2021; Leone et al., 2020) . Since we get more REM sleep in the morning, it could be speculated that the pandemic-related rise time delay may have increased the number of REM sleep morning awakenings, in turn enhancing the likelihood to remember vivid and intense dreams upon awakening. Indeed, albeit the view of an exclusive relationship between REM sleep and dreaming has been overcome (Scarpelli et al., 2022) , many studies point to more frequent vivid and emotional dream contents upon REM awakenings (Foulkes et al., 1988; Nielsen et al., 1991; Merritt et al., 1994; Hobson et al., 2000; Oudiette et al., 2012) . At present, only one study found that changes in sleep habits during the pandemic, including morning awakening, were associated with greater dream emotional intensity, more negative dream emotions, nightmare frequency and distress (Scarpelli et al., 2021a) . Therefore, the general hypothesis of a relationship between delayed rise time and changes in dream phenomenology during the pandemic needs of a further and direct assessment.

Finally, the present literature confirms during the pandemic the classical observation of a relationship between dream recall and demographic variables like gender and age. The J o u r n a l P r e -p r o o f existence of greater dream and nightmare frequency in females has been previously found (Nielsen et al., 2000; Nielsen, 2012; Schredl, 2010; Schredl and Reinhard, 2008) . The nature of this phenomenon is not clear, but the greater predisposition for depression, anxiety, insomnia, and larger emotional reaction to negative stimuli in women may play a role (Suh et al., 2018; Ozdin et al., 2020; Bradly et al., 2001; Lithari et al., 2010; Stevens and Hamann, 2012) . Concerning age, beyond some conflicting results in specific dream variables, a greater dream/nightmare frequency and intensity in younger individuals has been observed (Gorgoni et al., 2021; Frankl et al., 2021; Sommantico et al., 2021; Guerrero-Gomez et al., 2021; Musse et al. 2020; Scarpelli et al., 2021a; Scarpelli et al., 2022b) . This result is consistent with the drop in dream and nightmare recall frequency observed in previous studies (Scarpelli et al., 2019b) , likely explained by changes in sleep physiology (Nielsen, 2012) or reduced dream salience associated with lower interest in dreaming and impact of dream contents (Cohen, 1979; Giambra et al., 1996) .

The oneiric activity has been strongly affected by the pandemic. The present literature mainly points to a pandemic-related enhancement of dream and nightmare frequency, emotional intensity, and distressing contents, modulated by changes in the restrictive measures and associated with diurnal experiences, emotional status, and sleep pattern. We have highlighted several methodological problems of the reviewed studies that strongly Further studies are needed to support the present conclusions on the oneiric activity during the pandemic. Starting from the strong methodological heterogeneity observed, future efforts should be focused on the replication of the present findings. It is even more true for findings on lucid dreams and for the assessment of dreaming in specific populations. In particular, no conclusion can be drawn for results on the oneiric activity in healthcare workers, narcoleptic patients, and COVID-19 patients since they have been directly assessed only in single studies. Similarly, albeit lucid dreams have been investigated in 7 studies, their methods, objectives, and population of interest were heterogeneous, making it difficult to support specific hypotheses about this research topic.

Clearly, the unique condition represented by the lockdown during the first COVID-19 pandemic wave can't be replicated. Therefore, the replication of the available findings should be conducted through re-analysis of previously collected data. Moreover, a greater number of longitudinal studies on large samples in different phases of the pandemic is needed, with the aim to control the effect of environmental changes on dreams and nightmares. Finally, it is worth noting that the assessment of a possible publication bias concerning the reviewed literature is still missing. Therefore, in order to increase our understanding of pandemic-related changes in the oneiric activity, we strongly encourage the direct evaluation of the file drawer problem and reports of negative findings in this research field. 

-The COVID-19 pandemic induced an increase of dream frequency, intensity, and distressing contents -The phenomenology of the oneiric activity appears modulated by changes in restrictive measures -Diurnal experiences, emotional status, sleep, and trait factors affected pandemic dreaming -The literature on pandemic dreams is characterized by many methodological issues and large heterogeneity

Abbreviations: BDI-II

Nordic Sleep Questionnaire; DASS-21, Depression Anxiety Stress Scale-21

Third Edition of the International Classification of Sleep Disorders; ISI, Insomnia Severity Index; LIWC, Linguistic Inquiry and Word Count; PANSS, Positive and Negative Syndrome Scale; MADRE, Mannheim Dreams questionnaire; MOS-SS, Medical Outocomes Study-Sleep Scale; PHQ, Patient Health Questionnaire; PSQI

PTSD, Post-traumatic Stress Disorder

Sleep Behavior Disorder