Does sleep promote the generalization
across faces
Abstract
Episodic memory is a memory system for autobiographical events and its spatial temporal relations for a single event. Semantic memory is a memory system of general knowledge of the world. This general knowledge is formed by the extraction of statistical regularities across multiple events. This process is called generalization. In this study we are interested whether generalization across faces takes place and whether this process is facilitated by sleep. We are also interested whether we can replicate previous findings that sleep improves episodic memory. A memory task was used in which the subjects had to learn face-location associations on a board (64 black and white alternating squares; chessboard). Certain combinations of facial features were more prevalent in a corner. Every corner had a unique combination of facial features that was more prevalent. We are interested whether subjects extract this rule or regularity that is built in this task. The sleep group, slept the night after encoding, however, the sleep-deprived group was awake the whole the night after encoding. After three days recall took place. The sleep group and the sleep-deprived did not differ significantly on the episodic memory performance and the amount of generalization.
However, numerically there was a trend that the sleep group performed slightly better than the sleep-deprived group on the episodic memory performance. Moreover, numerically there was a trend in the generalization data as well. The sleep group generalized more than the sleep-deprived group.
Student Steffie Szegedi
Student number 5781892
Co-Assessor
Introduction
Memories make humans capable of learning from past events. This guides them in their decision making and makes them able to adapt to changing environmental demands (Rasch & Born, 2013). Although, memory is very important for functioning in everyday life, it does has its flaws in accuracy. These flaws were apparent in eyewitness testimony studies (Cutler et al., 1987) and in other studies as well, since phrasing of the question, suggestion (Loftus & Palmer, 1974) and misinformation (Ayers & Reder, 1998) alter the recall of an event. Although, the idea that the confidence of a memory is related to the accuracy of a recall seems reasonable, prior research has shown that positive negative and non-associations have been found between the confidence of a memory and its accuracy (Cutler et al., 1987; Desoto et al., 2014; Morgan & Southwick 2014). Moreover, a recent study found that eyewitness mistakes were responsible for the majority of the exonerations (Innocence Project 2011; O’Neill Schermer et al., 2011). Despite all this evidence the law system still depends heavily on eye witness testimonies. Therefore, it is relevant to provide more insight in how memories might change over time.
There are roughly two types of declarative memory: episodic memory and semantic memory. Episodic memory is described as a memory system for (autobiographical) events and its spatial-temporal relations (Moscovtich et al., 2006; Tulving, 1984) for a single episode (Battaglia & Pennartz, 2011). On the contrary semantic memory, also known as factual memory, is a memory system of general knowledge of the world (Moscovitch et al., 2006; Patterson et al., 2007). This general knowledge, is formed by the extraction of statistical regularities across multiple episodes (Battaglia & Pennartz, 2011) and is therefore context free (Vargha-Khadem et al., 1997; Patterson et al., 2007). The crucial role of the hippocampus in memory has been well established. However, to what extent the hippocampus is important per type of memory (semantic and episodic memory) and over time (recent vs remote memories) is still heavily debated. The possible role of the hippocampus in episodic and semantic memory is described in four different theories.
In the first theory, standard consolidation theory, the hippocampus is a temporary storage place for both episodic and semantic memories. By the process of consolidation the memory is transferred to the neocortex and the contribution of the hippocampus diminishes over time
(Moscovitch et al., 2006; Squire & Alvarez, 1995; Takashima et al., 2009; Winocur & Moscovitch, 2011).
In the second theory, the cognitive map theory, the hippocampus is important for making allocentric representations of the environment (episodic memory). The hippocampus is important for the retention and retrieval of both recent and remote episodic memories. However, semantic memory is independent of the hippocampus (Moscovitch et al., 2006; O’Keefe & Nadel, 1978).
In the third theory, the multiple trace theory (MTT), episodic memories, which are very detailed and vivid, depend on the hippocampus for both remote and recent memories. In this theory, semantic memory depends on the hippocampus for a limited period of time and then relies solely on the neorcortex (Moscovitch et al., 2006).
In the fourth theory, which has its roots in the MTT, episodic memories can also lose their detail and vividness and the general rule remains, so over time it becomes more similar to semantic memory (Moscovitch et al., 2006; Takashima et al., 2009; Winocur et al., 2007; Winocur, 2011). It has been proposed that hippocampal-neocortical communication is important for this extraction of statistical regularities (general rule) from episodes (Battaglia & Pennartz, 2011; McClelland et al., 1995). This process is called generalization and is also referred to as the transformation theory (Winocur & Moscovitch, 2011).
A process known to be important in memory (consolidation) is sleep (Diekelmann & Born, 2010; Maquet, 2001; Pennartz et al., 2002; Plihal & Born, 1997; Stickgold, 2005). More precisely, procedural memory is thought to benefit from REM sleep and declarative memory is thought to benefit from slow wave sleep (SWS) (Plihal & Born, 1997). It has been proposed that during sleep, assimilation and generalization of individual memory traces takes place, in which the exact representation of an episodic memory decays and the generalized meaning of gathered episodic memories remains (Walker & Stickgold, 2010).
In this study we are interested whether generalization across faces takes place and whether this process is facilitated by sleep. We are also interested whether we can replicate previous findings that sleep improves episodic memory. In our task, the subjects have to learn face-location associations on a screen. These face-face-locations associations are crucial to ensure the involvement of the hippocampus. This is important, because as previously stated, the communication between the hippocampus and neocortex is said to be important for the
extraction of regularities (generalization) across episodes. Moreover, humans with hippocampal damage have severe spatial memory impairments (Astur et al., 2002), but (face) recognition is not attenuated by a damaged hippocampus (Aggleton et al., 2005; Mayes et al., 2002; Turriziani et al., 2004; Varga-Khadem et al., 1997; Yonelinas et al., 2002).
Although, 1) prior research has shown that 10.5 hours of sleep is sufficient to recover from acute total sleep deprivation (Adam et al., 2006) and 2) a sleep protocol will be used to get the wake and sleep rhythm of the subjects back to normal as quickly as possible, we will conduct the PEBL perceptual vigilance task (PPVT) as a control measure. We are interested whether sleep deprivation the night after the learning phase of the task affects the consolidation of that information. This may lead to a diminished performance during the recall phase of the task. However, we want to rule out the possibility that the diminished performance is caused by exhaustion on the day of recall, because, it has been well established that sleep deprivation attenuates vigilance (Alhola & Polo-Kantola, 2007).
We hypothesize that sleep promotes generalization and improves memory. So the subjects in the sleep group are expected to make less mistakes than the sleep-deprived group in memorizing the faces. We also expect the subjects in the sleep group place the faces more often towards the prototype face than the sleep-deprived group.
Methods
Subjects
Of the 56 subjects that were recruited 14 were excluded. Exclusion criteria were: a) suboptimal general health due to infections, tiredness or allergies; b) sleeping problems, psychological, psychiatric and or neurological disorders; c) diabetes (1); d) heart diseases; e) insufficient performance on the dummy task, this task will be explained in more detail in the dummy task section (6); f) subjects that had participated in another study which used similar faces (6). And one subject did not show up. The remaining 42 were either assigned to the sleep condition or the sleep-deprived condition. The age range in the sleep condition group was between 18-27 years (M = 21.06, SD =2.437, N =18) all of them had a university level of education. In the sleep-deprived condition group the age range was between 19-25 years (M = 21.46, SD = 2.064, N = 24), 23 of them had a university level of education and one a college level. The subjects received study credits or money for completing the study. The study was approved by the Committee of Ethics of the University of Amsterdam. Written informed
consent was obtained from all subjects.
Procedure
The subjects performed three tasks: a) the dummy task, b) the prototype face task and c) the PEBL perceptual vigilance task (PPVT). Furthermore, the subjects also filled out four questionnaires: a) a general health and sleeping problems screener b) the Pittsburgh Sleep Quality Index, c) the Stanford Sleepiness Scale (SSS) and d) the sleep quality questionnaire.
The Dummy task
The dummy task was designed to make sure that the subjects had a quite good episodic memory, since the experimental task was difficult. The subjects were instructed to learn object-location associations (at which square the object was presented). In this task 20 pictures of everyday objects (e.g. chair, pizza, umbrella and dumbbells) were used. The total duration of the dummy task was 12 minutes. The objects were presented on a screen with a layout of 64 black and white squares alternating each other (similar to a chessboard; see figure 1). On 20 of the 64 squares an object was presented for 1 second (see the left picture of figure 1), then popped out of the square and became bigger, stayed big for 1 second (see the right picture of figure 1) and then became smaller again until it was the same size as the square again for 1 second (see the left picture of figure 1). Then after 3 seconds the next picture was presented at another square (location). After all the 20 pictures were shown, the subjects had a 30 second break and then round two started. Round 2 and 3 were exactly the same as round 1; the same pictures were shown in the exact same order. There was a 5 minute break between the three rounds of encoding and recall. During these 5 minutes the subjects were instructed to
solve a Sudoku puzzle.
Figure 1. The dummy task during encoding. Left picture is the original size, right picture is the popped out/ zoomed in size
During recall an object (one of the 20 that were used during encoding) was presented right from the chessboard (see figure 2) and then the subjects had to navigate the red square towards the location they thought the items were presented before (the order of presentation during recall was random). The navigation keys were T(↑) to go one square up, F(←) to go one square left, G(→) to go one square right, and V(↓) to go one square down on the computer keyboard; (direction arrows were placed on the cases with tape). When the red square was at the desired location the subjects had to press enter. They did this for all 20 pictures. The performance on the dummy task was sufficient when 1) each face was
misplaced 1 square or less, 2) 13 of the 20 faces were placed correctly at the exact same location as it was shown during encoding.
Figure 2. The dummy task during recall. Subjects had to navigate the red square to the location they thought the picture right from the chessboard was shown during encoding.
The prototype face task
The items that were used in the task were 64 gray scaled pictures of emotionally neutral faces produced in the program FacesTM software (IQ Biomertrix, 2003). Each face was composed of 6 features: a) age (either young or old adults); b) body composition (slender or stout); c) hair color (dark or light; only two colors due to gray scale); d) gender (women or men); e) headpiece (yes or no ); f) mole (yes or no) and g) glasses (yes or no). Every prototype category had a unique combination of two features, which was not present in the other prototype categories and each prototype category consisted of 16 unique faces. Prototype category A faces consisted of slender men, prototype category B faces consisted of elderly women, prototype category C faces consisted of blond haired young people and prototype category D faces consisted of dark haired stout people. The subjects were instructed to learn face-location associations (at which square the faces was presented). All the faces were presented at a particular square for 1 second (see left picture of figure 3), then popped out, this lasted for 1 second (see the right picture of figure 3) and then became smaller again back to
the size of the square for 1 second (see left picture again of figure 3) and then disappeared. Then after 3 seconds the next face was presented at another square (location). After all the 64 pictures were shown, the subjects had a 30 second break and then round two started. Round 2 and 3 were exactly the same as round 1; the same pictures were shown in the exact same order. The total duration of the prototype faces task was 30 minutes.
Figure 3. The prototype face task during encoding. Left picture is the original size, right picture is the popped out/zoomed in size.
In this task there were 64 black and white squares alternating each other (8 by 8; like a chessboard). Figure 4 presents a visualization of the task. In this task every corner had a prototype location. This prototype location is referred to as Pro I, Pro II, Pro III or Pro IV which is shown in Figure 4. The four prototype face categories were coupled to a prototype location (for example the prototype category A faces were coupled to prototype location I, and the prototype category B faces to location IV; this was randomized across subjects). The sixteen faces that belonged to a certain prototype category were placed around the board by a specific proportion. On the prototype locations there was always a prototype face that was coupled to that location. For example if prototype category A faces were coupled to prototype location I, then a prototype A face was placed on the prototype I location (100%). One square next to the prototype I location, there were 5 of the 8 squares filled with prototype category A faces (62.5%) and 3 of them by either prototype category B, C, or D faces (37.5%). Two squares next to the prototype I location there were 5 of the 11 squares filled with prototype category A faces (45.45%) and the remaining 6 with either prototype category B, C, or D faces (54.55%). Three squares next to the prototype location I, 3 of the 10 squares were filled
with prototype category A faces (30%) and the remaining 7 with either prototype category B, C, and D (70%). Four Squares next to the prototype there was 1 of the 10 squares filled with a prototype category A face (10%) and the remaining 9 with prototype category B, C and D faces (90%). Five Squares next to the prototype 1 of the 13 squares was filled with a prototype category A face (7.7%) and the remaining 12 squares with prototype category B, C and D (92.3%). This was also done for the other three prototypes according to the same proportion.
Figure 4. Prototype locations
During recall a certain object was presented right from the chessboard (see figure 5) and then the subjects had to navigate the red square towards the location they thought the items were presented before. The navigation keys were T (↑) to go one square up, F(←) to go one square left, G(→) to go one square right, and V(↓) to go one square down on the computer keyboard; (direction arrows were placed on the cases with tape). When the red square was at the desired location the subjects had to press enter. After this they had to indicate how confident they were of their response on a 5 point scale (1 (very unsure), 2 (unsure), 3 (indifferent), 4 (sure)
and 5 (very sure)). They did this for all 64 faces. The confidence rating variable is “reversed cumulative” (i.e. that a confidence rating of 1 and higher consisted of all the faces, a confidence rating of 2 and higher consisted all the faces with a confidence rating of 2 and higher (so 1 is not included) and a confidence rating of 3 and higher consisted of all the faces with a confidence rating of 3 and above (so faces rated with a confidence of 1 or 2 were not included) and so on.
Figure 5. The prototype face task during recall. Subjects had to navigate
the red square to the location they thought the picture right from the chessboard was shown during encoding.
A couple of variables were of interest in this task. The episodic memory performance was assessed by four groups of variables. The first group of variables, the number of correctly placed faces in total and for each prototype, was counted for every subject. The second group of variables, the proportion of correctly placed faces in total and for each confidence rating, was constructed to explore the possible relation between accuracy and confidence during recall. I have chosen for a proportion correctly placed faces, because the absolute number was not very informative, since there were not many faces with high confidence ratings (“reversed cumulative”). This means that if there were only 5 faces with a confidence rating of 5, but 4 of them were correct (80%) it seems absolutely a bad score but relatively it is quite good. For
example this subject placed 15 faces correctly with a confidence rating of 1 or higher it seems absolutely a good score but relatively it is quite poor (23.44%). The third group of variables, the episodic misplacement in total and for each prototype, was constructed as follows: the distance between the x coordinates of the encoding and recall coordinates divided by 80 (the length of one square was 80 pixels) squared plus the distance between the y coordinates of the encoding and recall coordinates divided by 80 squared. Then the square root of that answer was the misplacement per face (Pythagoras formula). √ (((X encoding - X recall)/80)²+ ((Y
encoding - Y recall)/80)²). This misplacement was calculated for all the 64 faces and the sum
of this misplacement was the total misplacement per subject. The misplacement per prototype was also calculated (this was done over 16 faces belonging to the same prototype category) and was the misplacement per prototype. The fourth group of variables, the misplacement in total and for each confidence rating, were constructed to explore the possible relation between the accuracy and confidence during recall. The misplacement in total was summed over all the faces that belonged to a certain confidence rating; for a confidence rating of 1 and greater the misplacement of all the faces were summed, for a confidence rating of 2 and greater the misplacement of the faces with a confidence rating of 2 and greater were summed and so on.
The generalization was assessed by two groups of variables. It was hypothesized that if generalization takes place, a face that belonged to a certain prototype category would be placed closer to the prototype location (it was coupled to; every prototype category was coupled to a certain prototype location) during recall than it was shown during encoding. Generalization was calculated as follows: the distance between encoding location and prototype location minus the distance between recall placement and prototype location. This can be calculated by the following formula: (√ (((X prototype - X encoding)/80)²+ ((Y
prototype - Y encoding/80)²)) – (√ (((X prototype - X recall)/80)²+ ((Y prototype - Y recall)/80)²)), so this implies the following situations:
1) (√ (((X prototype - X encoding)/80)²+ ((Y prototype - Y encoding/80)²)) - (√ (((X prototype - X recall)/80)²+ ((Y prototype - Y recall)/80)²)) > 0 means generalization and episodic misplacement.
2) (√ (((X prototype - X encoding)/80)² + ((Y prototype - Y encoding/80)²)) - ( √ (((X prototype - X recall)/80)² + ((Y prototype - Y recall)/80)²)) = 0 means either placing the face exactly correct or in some random instances not correct but at the exact same distance in another direction and no generalization
3) (√ (((X prototype- X encoding)/80)² + ((Y prototype - Y encoding/80)²)) - (√ (((X prototype - X recall)/80)² + ((Y prototype - Y recall)/80)²)) < 0 means, no generalization and episodic misplacement.
The first group of variables was the generalization in total and for each prototype. The generalization was calculated for all the 64 faces and the sum of the generalization was the total generalization per subject. The generalization per prototype was also calculated (this was done over 16 faces belonging to the same prototype category) and was the generalization per prototype. The second group of variables, the generalization in total and for each confidence rating, was constructed to explore the possible relation between generalization and confidence during recall. This was of interest, because in order to generalize a subject had to make a mistake. There are two kind of “mistakes” possible in this task a generalization mistake and a random error mistake. We expected the faces with a lower confidence rating to contain more random error mistakes than the faces with a higher confidence rating. This random error or noise might blur the generalization outcome. The generalization for each confidence rating was calculated as follows: the generalization in total was summed over all the faces that belonged to a certain confidence rating; for a confidence rating of 1 and greater the generalization of all the faces were summed, for a confidence rating of 2 and greater the generalization of the faces with a confidence rating of 2 and greater were summed and so on.
Another measurement, which I will refer to as “generalizable”, was constructed to make it independent of episodic memory and is based on the generalization measurements. Both the correct prototype faces, the incorrect prototype faces and the correct faces could not be placed closer to the prototype during recall than it was shown during encoding, therefore generalization cannot occur. It is calculated as follows: “Generalizable” = total
generalization without prototypes / ((64 - number correct) - the number of erroneous placed prototypes). The “generalizable” is assessed by two different groups of variables. The
first group of variables, was the “generalizable” in total and for each prototype. The “generalizable” was calculated for all the 64 faces and the sum of the “generalizable” was the total “generalizable” per subject. The “generalizable” per prototype was also calculated (this was done over 16 faces belonging to the same prototype category) and was the “generalizable” per prototype. The second group of variables, the “generalizable” in total and for each confidence rating, was constructed to explore the possible relation between “generalizable” and confidence during recall. This was of interest, because in order to
generalize a subject had to make a mistake. There are two kind of “mistakes” possible in this task a generalization mistake and a random error mistake. We expected the faces with a lower confidence rating to contain more random error mistakes than the faces with a higher confidence rating. This random error or noise might blur the “generalizable” outcome. The “generalizable” for each confidence rating was calculated as follows: the “generalizable” in total was summed over all the faces that belonged to a certain confidence rating; for a confidence rating of 1 and greater the “generalizable” of all the faces were summed, for a confidence rating of 2 and greater the “generalizable” of the faces with a confidence rating of 2 and greater were summed and so on.
PEBL perceptual vigilance task (PPVT)
As a control measure the PPVT was used, which is a sustained attention task, to control for possible differences between the sleep and sleep-deprived groups on sleepiness and or tiredness during encoding and recall. During encoding and retrieval the subjects in the sleep and the sleep-deprived condition should not differ on sleepiness or tiredness. In this task subjects had to press as quickly as possible at the space bar when a white cross was presented and should not press the button when an orange circle was presented. The time between start sign and presenting the white cross differed and could be 1000, 2000, 3000, 4000, 5000, 6000, 7000, 8000 or 9000 milliseconds. The time between the appearance of the white cross and pressing the space bar is the reaction time of the subject. The reaction time is averaged over all the 120 trials per subject. Reaction times exceeding 500 milliseconds are called lapses and are counted. Reaction times exceeding 30 seconds are called sleep attacks and are counted as well. The output files of 11 subjects on day 1 and the output files of 10 subjects on day 4 were not complete: all the reaction times were present in the output file, but the sleep lapse and sleep attack information were missing.
Questionnaires
The Stanford Sleepiness Scale (SSS) consisted of 7 phrases which varied from active or alert towards drowsy or non-alert. Subjects had to choose which statement resembled them the most.
The sleep quality questionnaire consisted of 14 questions, which were related to how well the subjects slept the night before.
The Pittsburgh Sleep Quality Index was used to address sleeping problems and the quality of sleep in a very elaborate manner.
General health and sleeping problems screener (this was a self-made screener to make sure that all the subjects were in good health and had no sleeping problems, since sleep deprivation of 8 hours may have strong effects on both the physical and psychological wellbeing of human beings).
We used another home-made questionnaire to determine whether subjects noticed anything in particular in the task, or had any clues about a general rule in the task. See appendix I for all the questionnaires.
Protocol
The general health questionnaire, known sleeping problems and known problems with sleep deprivation were addressed beforehand by email.
The subjects were expected at 8:00 pm at the sleep labs of the University of Amsterdam. They filled out the informed consent, the approval form to contact their doctor, the Stanford Sleepiness Scale, the sleep quality questionnaire and the Pittsburgh sleep quality index.
Then the encoding of the dummy task took place; between encoding and recall they made a Sudoku for 5 minutes. After the recall part of the task, the subjects had a break for 15 minutes. If their performance was insufficient they were sent home and did not participate in the rest of the study. The encoding part of the prototype face task took place after this, then the subjects had another break for about 15 minutes and then they did the PPVT. After the PPVT the subjects had some time off; they were allowed to do some reading in magazines, books or play with their cell phone.
The subjects in the sleep group prepared for bed, then the EEG apparatus was attached to the scalp of the subjects. At 11 pm it was bed time. In the sleep-deprived group the subjects had a whole night of free time in which they were allowed to: eat, drink (no caffeine containing drinks), play board games and watch animal movies, (no films with human faces were allowed to watch). Falling asleep was prohibited. This was checked by a supervisor who was present and awake the whole night.
In the sleep condition the 64-channel EEG caps were used. Three different cap sizes were used. The cap size is important because if too lose, the channels cannot get a good EEG signal
from the scalp, but if too tight it is uncomfortable for the subject. The skin of the forehead, behind the ear and the chin had to be cleaned with alcohol in order to get a good signal for the EOG and EMG electrodes (placed on the chin). Each channel had to be gelled. The impedance should be as low as possible, especially the CZ which was the ground. Because if that channel its impedance is bad, the impedances of all the other channels are bad as well. The settings in ASA were Wave guard 64-64 amplifier or 64-128 amplifier (depending on which amplifier was used; we had a 64 channels Refa amplifier of the Twente Medical system and a 128 channels Refa amplifier of the Twente Medical System), sample rate 512 Hz. Wake time was at 8:00 am, the EEG cap got removed. The subjects filled out the Stanford sleepiness scale and the sleep quality questionnaire. The subjects were instructed to go to bed between 10:00 pm and 12:00 pm and to get up between 7:00 am and 9:00 am the days between encoding and recall (day 1 and day 4). Then they left at approximately 8:30 am.
The sleep-deprived subjects filled out the Stanford sleepiness scale at 8:00 am. The other supervisor (which was not present during the sleep deprivation night) checked whether the subjects were feeling all right and if they were able to go home on their own. Between 12:00-3:00 pm the subjects had to sleep (or try to sleep) for about 90 minutes. And were not allowed to sleep during the rest of the day. The subjects were instructed to start their night sleep between 10:00 pm and 12:00 pm and to have a full night of sleep (7-9 hours). This sleep protocol was used, to get their wake and sleep rhythm back to normal as quick as possible. They left at approximately 8:30 am.
On the fourth day the subjects of both groups had to come back at 10:00 am for the recall part of the task, which took around 20 minutes (it depended on how quickly a subject chooses an answer). Afterwards they did the PPVT again and filled out the Stanford Sleepiness Scale and the sleep quality questionnaire. They left the laboratory around 10:40 am.
Analysis
Outlier detection was done for all the variables, data points with a z score of 3.29 or greater were removed from the analysis. Some variables met the criteria for parametric testing after the removal of (an) outlier(s). In these cases a parametric test was performed.
For comparing the sleep group and the sleep-deprived group (between subjects design) an independent t-test was performed, on all the data that met the criteria for parametric testing.
For the variables that did not meet the criteria of parametric testing the Mann-Whitney U test was performed (excact).
A within subjects design was used for comparing the sleepiness and alertness questionnaires and PPVT at the day of encoding and at the day of recall. For the variables that met the criteria of parametric testing the dependent t-test was performed and for the variables that did not meet the criteria the Wilcoxon signed-rank test (exact) was used.
Results
Sleep and alertness questionnaires and the PPVT
The alertness and sleep questionnaires and the PPVT were conducted to control for possible differences in alertness and sleepiness between the sleep group and the sleep-deprived group on the day of encoding and recall. Since, attenuated alertness during the learning part of the prototype face task might lead to poorer encoding of the information. And attenuated alertness during recall might lead to more errors or poorer retrieval. The PPVT variables consisted of the average reaction time, the number of lapses and the number of sleep attacks per subject.
None of the subjects had a sleep attack, which is a reaction time exceeding 30 seconds. As expected, the sleep group and the sleep-deprived group did not differ significantly on all the sleep and attention variables on the day of encoding. At day two, which is the morning after a night of sleep for the sleep group and a night without sleep for the sleep-deprived group, the subjects filled out the SSS. As expected, there was a significant difference between the sleep group (Mdn = 3.0, Q1 = 2.0, Q3 = 4.0) and the sleep deprivation group (Mdn = 4.0, Q1 = 3.5, Q3 = 5.0) on the SSS (U = 96.5, p = 0.002), which indicates that the sleep-deprived group felt less alert than the sleep group at day 2. As expected, the sleep group and the sleep-deprived group did not differ significantly on almost all the sleep and attention variables on the day of recall. However, the sleep group (Mdn = 12.0, Q1 = 12.0, Q3 = 14.0) had a better quality of sleep than the sleep-deprived group (Mdn = 11.0, Q1 = 8.5, Q3 = 12.0), (U = 110.5, p = 0.009) the night before recall. See table 1 for all the sleep and alertness statistics.
Table 1. Statistics table of the sleep questionnaires and the PPVT. The number behind a variable indicated at which day the measurement took place. Day 1 is the day of encoding, day 2 the morning after a night of sleep or sleep deprivation, day 4 is the day of recall. Variables denoted with an asterisk* were not normally distributed and the median was used instead of the mean and the Mann-Whitney U test instead of the
Independent T-test. M: Mean, Mdn: Median, SD: Standard Deviation, Q1: first Quartile, Q3: third Quartile, N: sample size, T: independent t-test, U: Mann-Whitney U test, df: degrees of freedom and p: p-value.
Episodic memory
As a first measure of episodic memory, the number of correctly placed faces was compared between the sleep group and the sleep-deprived group. This was done for all the correctly placed faces in total and per prototype. The sleep group performed slightly better than the sleep-deprived group on these variables, but it was far from significant. Table 2 presents an overview of the statistics and figure 6 is a visualization of the results.
Table 2. Statistics table of the average number of correctly placed faces per group in total and for each prototype. The variables denoted with an asterisk* were not normally distributed and the median was used instead of the mean and the Mann-Whitney test instead of the independent t-test. M: Mean, Mdn: Median, SD: Standard Deviation, Q1: first Quartile, Q3: third Quartile, N: sample size, T: independent t-test, U: Mann-Whitney U test, df: degrees of freedom and p: p-value.
Figure 6. The average number of correctly placed faces per group in total and for each prototype.
As a second measure of episodic memory, the proportion of correctly placed faces per confidence rating was compared between the sleep group and the sleep-deprived group. We were interested whether higher confidence is related to a higher performance. A proportion of correctly placed faces was used, because the absolute number of correctly placed faces
decreases when a higher confidence rating or greater was used (not so many faces fall in the group with a confidence rating of 5, but all the 64 faces fall in the category of a confidence rating of 1 and greater), but relatively the number of correctly placed faces increases see table 3 and figure 7.
Table 3. Statistics table of the average proportion of correctly placed faces per group in total and for each confidence rating. The variables denoted with an asterisk* were not normally distributed and the median was used instead of the mean and the Mann-Whitney test instead of the independent t-test. M: Mean, Mdn: Median, SD: Standard Deviation, Q1: first Quartile, Q3: third Quartile, N: sample size, T: independent t-test, U: Mann-Whitney U test, df: degrees of freedom and p: p-value.
Figure 7. The average proportion of total correctly placed faces per group and for each confidence rating.
As a third measure of episodic memory, the episodic misplacement was compared between the sleep group and the sleep-deprived group. This is the distance between the location where the face was shown during encoding and where the face was placed by the subject during recall. The misplacement of the faces during recall was bigger for the sleep-deprived group than for the sleep group in total and per prototype, however contrary to our hypothesis none of the differences were statistically significant see table 4 and figure 8.
Table 4. Statistics table of the average episodic misplacement per group in total and for each prototype. The variables denoted with an asterisk* were not normally distributed and the median was used instead of the mean and the Mann-Whitney test instead of the independent t-test. M: Mean, Mdn: Median, SD: Standard Deviation, Q1: first Quartile, Q3: third Quartile, N: sample size, T: independent t-test, U: Mann-Whitney U test, df: degrees of freedom and p: p-value.
Figure 8. The average misplacement of the faces per group in total and for each prototype.
As a fourth measure of episodic memory, the episodic misplacement in total and for each confidence rating was compared between the sleep group and the sleep-deprived group. We were interested whether the confidence of an answer during recall is related to the accuracy during recall. With an increasing confidence rating the misplacement decreases. However, it should be borne in mind that the confidence rating variable is “reversed cumulative” (i.e. that a confidence rating of 1 and higher consisted of all the faces, a confidence rating of 2 and higher consisted all the faces with a confidence rating of 2 and higher (so 1 is not included) and a confidence rating of 3 and higher consisted of all the faces with a confidence rating of 3 and above (so faces rated with a confidence of 1 or 2 were not included) and so on. This implies that the decrease in misplacement distance is not only caused by a higher confidence, but is also influenced by the decrease in number of faces falling in the next confidence rating category. The results of the episodic misplacement for each confidence rating were mixed when the sleep group and the sleep-deprived group were compared see table 5 and figure 9. The differences were very small. It might seem that the results in table 5 do not match the results in figure 9 regarding the confidence rating of 2 and higher, 4 and higher and 5 and higher. This is because in the figure the mean is used for all the variables and in the table the median is used of the not normally distributed data and the mean for the normally distributed data.
Table 5. Statistics table of the average episodic misplacement per group in total and for each confidence rating. The variables denoted with an asterisk* were not normally distributed and the median was used instead of the mean and the Mann-Whitney test instead of the independent t-test. M: Mean, Mdn: Median, SD: Standard Deviation, Q1: first Quartile, Q3: third Quartile, N: sample size, T: independent t-test, U: Mann-Whitney U test, df: degrees of freedom and p: p-value.
Figure 9. The average misplacement of the faces per group in total and for each confidence rating.
Generalization
As a first measure of generalization, the generalization in total and for each prototype was compared between the sleep group and the sleep-deprived group. In general the sleep group showed more generalization than the sleep-deprived group, but it was not significant. However, for the prototype A faces the sleep group (M = -5.7, SD = 6.2) showed more generalization than the sleep-deprived group (M = 9.9, SD = 6.6) t(39) = 2.1, p = 0.043. For the prototype D faces the sleep group (M = -10.1, SD = 6.5) showed more generalization than the sleep-deprived group (M = -15.3, SD = 8.3) as well t(40) = 2.2, p = 0.035 see table 6 and figure 10 for more details.
Table 6. Statistics table of the average generalization per group in total and for each prototype. The variables denoted with an asterisk* were not normally distributed and the median was used instead of the mean and the Mann-Whitney test instead of the independent t-test. M: Mean, Mdn: Median, SD: Standard Deviation, Q1: first Quartile, Q3: third Quartile, N: sample size, T: independent t-test, U: Mann-Whitney U test, df: degrees of freedom and p: p-value.
Figure 10. The average generalization across faces per group in total and for each prototype.
As a second measure of generalization, the generalization in total and for each confidence rating was compared between the sleep group and the sleep-deprived group. We were interested whether the confidence of an answer during recall was related to generalization. This was of interest, because in order to generalize a subject had to make a mistake. There are two kind of “mistakes” possible in this task a generalization mistake and a random error mistake. We expected the faces with a lower confidence rating to contain more random error mistakes than the faces with a higher confidence rating. This random error or noise might blur the generalization outcome.
There were no significant differences between the sleep group and the sleep-deprived group on any of the variables see table 7 and figure 11. The generalization increased with an
increased confidence rating, however because of the “reversed cumulative” nature of the confidence rating variables, this increase is not only caused by the increased confidence. With an increasing confidence rating the differences between the sleep group and the sleep-deprived group did not become more evident. It was expected to become more evident, because the faces with a higher confidence rating were expected to contain less noise.
Table 7. Statistics table of the average generalization per group in total and for each confidence rating. The variables denoted with an asterisk* were not normally distributed and the median was used instead of the mean and the Mann-Whitney test instead of the independent t-test. M: Mean, Mdn: Median, SD: Standard Deviation, Q1: first Quartile, Q3: third Quartile, N: sample size, T: independent t-test, U: Mann-Whitney U test, df: degrees of freedom and p: p-value.
Figure 11. The average generalization across faces per group in total and for each confidence rating.
As a first measure of “generalizable”, the “generalizable” in total and for each prototype was compared between the sleep group and the sleep-deprived group. This variable was created to be as independent as possible of the episodic memory measurements. In order to generalize subjects need to make mistakes, hence correctly placed faces have 0 generalization. However, in the case of the “generalizable” variable, only the faces that could be generalized were taken into account. The prototype faces and the correctly placed faces cannot be generalized and therefore were not included in this measurement. The sleep group performed slightly better on the “generalizable” variable than the sleep-deprived group, but it was not significant see table 8 and figure 10 for the results.
Table 8. Statistics table of the average “generalizable” per group in total and for each prototype. The variables denoted with an asterisk* were not normally distributed and the median was used instead of the mean and the Mann-Whitney test instead of the independent t-test. M: Mean, Mdn: Median, SD: Standard Deviation, Q1: first Quartile, Q3: third Quartile, N: sample size, T: independent t-test, U: Mann-Whitney U test, df: degrees of freedom and p: p-value.
Figure 12. The average “generalizable” across faces per group in total and for each prototype.
As a second measure of “generalizable”, the “generalizable” in total and for each confidence rating was compared between the sleep group and the sleep-deprived group. We were
interested whether the confidence of an answer during recall is related to the “generalizable”. The sleep group performed slightly better than the sleep-deprived group, but it was not significant see table 9 and figure 11 for the results. With an increasing confidence rating the “generalizable” increased, however it should be borne in mind as well that the variable confidence rating is reversed accumulative; which implies that the increase of generalization is not merely caused by the increase in confidence. With an increasing confidence rating the differences between the sleep group and the sleep-deprived group did not become more evident. It was expected to become more evident, because the faces with a higher confidence rating were expected to contain less noise.
Table 9. Statistics table of the average “generalizable” per group in total and for each confidence rating. The variables denoted with an asterisk* were not normally distributed and the median was used instead of the mean and the Mann-Whitney test instead of the independent t-test. M: Mean, Mdn: Median, SD: Standard Deviation, Q1: first Quartile, Q3: third Quartile, N: sample size, T: independent t-test, U: Mann-Whitney U test, df: degrees of freedom and p: p-value.
Figure 13. The average “generalizable” across faces per group in total and for each confidence rating.
The current study addressed two research questions. The first question was, to investigate whether sleep compared to sleep deprivation the night after encoding improves episodic memory performance. The second question was whether sleep compared to sleep deprivation the night after encoding increases generalization.
The current study did not find a significant difference between the sleep and the sleep-deprived group on episodic memory performance. It was expected that the subjects in the sleep group would perform better than the sleep-deprived group, since prior studies have shown that sleep plays an important role in memory consolidation (Diekelmann & Born, 2010; Maquet, 2001; Pennartz et al., 2002; Plihal & Born, 1997; Stickgold, 2005). However, there is a numerical trend in the episodic memory data, that the sleep group performed better than the sleep-deprived group, but it did not reach significance.
It was also expected that the sleep group would show more generalization than the sleep-deprived group (Walker & Stickgold, 2010), but the results in this research do not support this view. However, there is a numerical trend in the generalization data, that the sleep group performed slightly better than the sleep-deprived group.
Limitations /Explanations
There are a couple of factors that might have influenced the results. The none significant results might be due to the small sample size, since smaller sample sizes have less statistical power. The task was quite difficult, so floor effects may account for the lack of differences found between the sleep group and the sleep-deprived group. We tried to anticipate on this task difficulty problem by doing an entry level task (dummy task).
The generalization measurement we used, was in general negative across subjects, which implies random error. This might be due to the difficulty of this task, that many mistakes were random errors, because the subjects had no clue where the face was shown during encoding. Although, with an increasing confidence rating the generalization increased it was still negative. A problem within this task, is the fact that generalization is erroneous in nature. Prototype category A faces that were shown far from the prototype location during encoding had a higher chance to be placed closer to the prototype location than prototype category A faces that were shown closer to the prototype location during encoding. This makes it difficult to distinguish between the generalization error and the random error. However, this problem
might be partly solved by using quadrants. In this study generalization was operationalized as follows: when a face was placed closer to their prototype location during recall than it was shown during encoding. Generalization is the extraction of statistical regularities (general rule) across episodes. The general rule in this task is that certain faces (for example the prototype category A faces) are shown (more frequently) in the upper left corner (it’s frequency decreases towards the other corners of the chessboard). Intuitively, a quadrant resembles this general rule quite well. Moreover, when quadrants are used, faces have a more similar chance to be generalized. However, with this method it is difficult to distinguish between episodic memory and generalization. These two measurements may overlap when a face is placed correctly, but it is also within the quadrant it belongs to. In that case there is both a correct episodic memory performance and generalization.
Another factor that might have influenced the results, is that sleep deprivation does not merely mean no sleep for a period of time. There are indications that unintentionally it also induces increased stress and changed arousal levels (Paller & Voss, 2004).
Future Directions
Although the idea behind this task is very interesting, some changes might improve this task a little. First, faces are very complex stimuli and certain features are more salient than others. Perhaps instead of using faces, shapes could be used. They are less complex, but still rules (categories) could be extracted. Secondly, as mentioned earlier quadrants could be used, because then a face could be generalized and placed correctly at the same time. Thirdly, learning 64 shapes or faces could be way too many, perhaps reducing this number of shapes or faces, decreases the random errors.
Insight in processes or circumstances that alter memories is of great importance in both everyday life and during court cases. Altering memories might even be an therapeutic outcome for war veterans or other sufferers of (severe) trauma’s.
References
Aggleton, J. P., Seralynne, D. V., Denby, C., Dix, S., Mayes, A. R., Roberts, N. Yonelinas, A. P. (2005). Sparing of the familiarity component of recognition memory in a patient with hippocampal pathology. Neuropsychologia, 43, 1810 – 1823.
Ayers, M. S. & Reder, M. L. (1998). A theoretical review of the misinformation effect: Predictions from an activation-based memory model. Psychonomic Bullentin & Review, 5, 1, 1 – 21.
Carr, M. F., Jadhav, S. P. & Frank, L. M. (2011). Hippocampal replay in the awake state: a potential substrate for memory consolidation and retrieval. Nature Neurosciene, 1, 2, 147 – 153.
Diekelmann, S. & Born, J. (2010). The memory function of sleep. Nature Reviews Neuroscience, 11, 114 – 126.
DeSoto, K. A. & Roediger III, H. L. (2014). Positive and Negative Correlations Between Confidence and Accuracy for the same Events in Recognition of Categorized Lists. Psychological Science, 1 – 8.
Loftus, E. F. & Palmer, J. C. (1974). Reconstruction of Automobile Destruction: An example of the Interaction Between Language and Memory. Journal of Verbal Learning and Verbal Behavior, 13, 585 – 589.
Maquet, P. (2001). The Role of Sleep in Learning and Memory. Science, 294, 1048 – 1052.
Morgan, C. A. & Southwick, S. (2014). Perspective: I believe what I remember, but it may not be true. Neurobiology of Learning and Memory, in press.
Mayes, A. R., Holdstock, J. S., Isaac, C. L., Hunkin, N. M. & Roberts, N. (2002). Relative Sparing of Item Recognition Memory in a Patient With Adult-Onset Damage Limited to the Hippocampus. Hippocampus, 12, 325 – 340.
Moscovitch, M., Nadel, L., Winocur, G., Gilboa, A. & Rosenbaum, R. S. (2006). Current Opinion in Neurobiology, 16, 179 – 190.
Paller, K. A. & Voss, J. L. (2004). Memory reactivation and consolidation during sleep. Learining and Memory, 11, 664 – 670.
Pennartz, C. M. A., Uylings, H. B. M., Barnes, C. A. & McNaughton, B. L. (2002). Memory reactivation and consolidation during sleep: from cellular mechanisms to human performance. Progress in Brain Research, 38, 143 – 166.
Plihal, W. & Born, J. (1997). Effects of Early and Late Nocturnal Sleep on Declarative and Procedural Memory. Journal of Cognitive Neuroscience, 9, 4, 534 – 547.
Stickgold, R. (2005). Sleep-dependent memory consolidation. Nature, 437, 1272 – 1278.
Takashima, A., Nieuwenhuis, I. L. C., Jensen, O., Talamini, L. M. Rijpkema, M. & Fernández, G. (2009). Shift from Hippocampal to Neocortical Centered Retrieval Network with Consolidation. The Journal of Neuroscience, 29, 32, 10087 – 10093.
Turriziani, P., Fadda, L., Caltagirone, C. & Carlesimo, G.A. (2004). Recognition memory for single items and for associations in amnesic patients. Neuropsychologia, 42, 426 – 433.
Varga-Khadem, F., Gadian, D. G., Watkins, K. E., Connelly, A., Van Paesschen, W. & Mishkin, M. (1997). Differential Effects Hippocampal Pathology on Episodic and Semantic Memory. Science, 277, 376 – 380.
Walker, M. P. & Stickgold, R. (2010). Overnight alchemy: Sleep-dependent memory evolution. Nature Neuroscience Correspondence.
Wells, G. L., Olson, E. A. & Charman, S. D. (2002). The Confidence of Eyewitnesses in Their Identifications From Lineups. Current Directions in Psychological Science, 11, 5, 151 – 154.
Winocur, G., Moscovitch, M. & Sekeres, M. (2007). Memory consolidation or transformation: context manipulation and hippocampal representations of memory. Nature Neuroscience, 10, 5, 555 – 557.
Winocur, G. & Moscovitch, M. (2011). Memory Transformation and Systems Consolidation. Journal of the International Neuropsychological Society, 17, 766 – 780.
Yonelinas, A. P., Kroll, N. E. A., Quamme, J. R., Lazzara, M. M., Sauvé, M., Widaman, K. F. & Knight, R. T. (2002). Effects of extensive temporal lobe damage or mild hypoxia on recollection and familiarity. Nature Neuroscience, 5, 11, 1236 – 1241.
Appendix 1
STANFORD SLEEPINESS SCALE
NAAM:
DATUM:
TIJD:
Geef aan welke van de volgende groepen van omschrijvingen het best weergeeft hoe u zich op dit moment voelt. Wilt u slechts één van de zeven groepen aankruisen?
[] aktief; energiek; alert; klaar wakker.
[] goed, maar niet maximaal funktionerend; in staat tot concentreren.
[] ontspannen; wakker; niet volledig alert; reagerend op wat er om me heen gebeurt
[] een beetje duf; niet honderd procent funktionerend; enigszins mat
[] duf; moeite om wakker te blijven; loom
[] slaperig; zin om te gaan liggen; vechtend tegen de slaap; wegdommelend.
[] zwevend tussen waken en slapen; op het punt in slaap te vallen; niet meer in staat tegen de slaap te vechten
Naam: Datum: ppncode:
Wilt U de nu volgende Slaapkwaliteits Schaal volledig invullen?
Het is de bedoeling dat U ofwel "eens" omcirkelt, ofwel "oneens", ook in die gevallen waar U nauwelijks kunt beslissen. Denkt U niet te lang na, omcirkelt U volgens Uw eerste indruk. Er is dus buiten "eens" en "oneens" geen alternatief mogelijk.
SLAAP KWALITEIT.
Ik vind dat ik vannacht heel slecht geslapen heb. Eens/Oneens
Ik lag gisteravond langer dan een half uur wakker voor ik insliep Eens/Oneens
Ik ben vannacht meerdere malen wakker geworden Eens/Oneens
Ik had vanochtend na het opstaan een moe gevoel Eens/Oneens
Ik ben vannacht naar mijn gevoel slaap tekort gekomen Eens/Oneens
Ik ben vannacht opgestaan Eens/Oneens
Ik voelde me vanochtend, nadat ik was opgestaan, goed uitgerust Eens/Oneens
Ik heb naar mijn gevoel vannacht maar een paar uur geslapen Eens/Oneens
Ik vind dat ik vannacht goed geslapen heb Eens/Oneens
Ik heb vannacht geen oog dicht gedaan Eens/Oneens
Ik sliep gisteravond gemakkelijk in Eens/Oneens
Ik had vannacht nadat ik wakker geworden was, moeite weer in slaap te vallen Eens/Oneens
Ik lag vannacht erg te woelen Eens/Oneens
De volgende vragen hebben alleen betrekking op uw slaap van de afgelopen maand. Uw antwoorden moeten een zo nauwkeurig mogelijke weergave zijn van de meerderheid van de dagen en nachten tijdens de afgelopen maand. Gelieve alle vragen te beantwoorden.
1. Hoe laat ging u 's avonds meestal naar bed gedurende de afgelopen maand?
2. Hoeveel minuten duurde het de afgelopen maand gewoonlijk voordat u in slaap viel, wanneer u wilde gaan slapen? Aantal minuten (per nacht): 3. Hoe laat stond u, gedurende de afgelopen maand, meestal op?
4. Aan hoeveel uren slaap kwam u gemiddeld per nacht tijdens de afgelopen maand?
(Dit aantal kan verschillen van het aantal dat u in bed doorbracht.) Aantal uren slaap per nacht: Geef bij de volgende vragen aan welk antwoord dat op u van toepassing is. Sla geen vraag over.
5. Hoe vaak had u tijdens de afgelopen maand moeilijkheden met slapen, omdat u...
niet tijdends de afgelopen maand minde r dan eenm aal per week een of tweem aal per week drie of meerm aal per week niet in slaap kon vallen binnen 30 minuten? 's nachts of in de vroege morgen wakker werd? naar het toilet moest gaan?
niet makkelijk kon ademhalen? luid hoestte of snurkte? het te koud had? het te warm had? nachtmerries had? pijn had? (een) andere reden(en) had?
zeer goed redelijk goed vrij slecht zeer slecht 6. Hoe zou u uw globale slaapkwaliteit tijdens de afgelopen maand beoordelen? niet tijdens de afgelopen maand minder dan eenmaal per week een of tweemaal per week drie of meermaal per week
7. Hoe vaak nam u
gedurende de afgelopen maand medicijnen (al dan
niet
voorgeschreven)?
Als u medicijnen
gebruikte was dat dan om beter te kunnen slapen?
Ja Nee
Als u voor andere redenen medicijnen
gebruikt zou u dat dan hier willen invullen?
niet tijdens de afgelopen maand minder dan eenmaal per week een of tweemaal per week drie of meermaal per week
8. Hoe vaak had u
het de voorbije maand moeilijk om wakker te blijven tijdens verschillende bezigheden, zoals eten of deelname aan sociale activiteiten?
geen enkel probleem
slechts een klein
probleem
enigzins een probleem
een heel groot probleem 9. In welke mate
was het de afgelopen maand voor u een probleem om voldoende
dagelijkse
activiteiten uit te voeren??
10. Indien u een kamergenoot of bedpartner heeft, vraag hem/haar hoe vaak u tijdens de afgelopen maand tijdens uw slaap...
(als u geen kamergenoot of bedpartner heeft kunt u overal niet van toepassing invullen)
niet tijdends de afgelopen maand minder dan eenmaal per week een of tweemaal per week drie of meermaal per week niet van toepassing luid snurkte lange ademhalingspauzes had trekkende of schoppende benen had periodes van verwardheid had een andere rusteloosheid had. omschrijf:
ochtendmens avondmens geen van beide
11. Wat voor type persoon bent u?
12. Wat is uw gewicht? kilo 13. Wat is uw lengte in centimeters? cm 14. Hoeveel glazen alcohol drinkt u gemiddeld per dag? glazen 15. Hoeveel cafeïne houdenede dranken drinkt u gemiddeld per dag? glazen
Vragenlijst slaap en geheugen
pp …….1. Zijn je bij het maken van deze taak bepaalde dingen opgevallen
2. Had je het idee dat bepaalde gezichten vaker voor kwamen op het memorybord, denk aan bijv oudere mannen, vrouwen met bril enz.
3. De taak was zo opgebouwd dat in elke hoek van het memorybord een bepaald type gezicht (de 4 categorieen) vaker voor kwamen. Was je dit opgevallen?
Intake vragenlijst
Algemene informatie Naam:Leeftijd:
Geslacht: Man / Vrouw
Links – of rechtshandigheid:
Opleidingsniveau: MBO/HBO/WO
Medische checklist
1. Heb je momenteel een psychologische/psychiatrische aandoening of een historie daarvan? Ja /Nee
2. Heb je momenteel epilepsie of dergelijke verschijnselen of een historie daarvan? Ja/Nee
Zo Ja wat voor?
3. Heb je momenteel andere psychologische/psychiatrische aandoeningen of een historie daarvan? Ja/Nee
4. Heb je cardiovasuclaire klachten en of aandoeningen? Ja / Nee Zo Ja welke?
5. Heb je Diabetes type I of II Ja / Nee
6. Ben je momenteel ziek of heb je klachten? Ja / Nee Zo Ja welke?
Slaap checklist
7. Heb je slaap problemen? Ja / Nee Zo Ja wat voor?
8. Heb je een meer dan gemiddelde gevoeligheid voor slaapverstoringen Ja / Nee Waarin uit zicht dat?
9. Heb je een normaal Slaapwaak ritme (Slaap tussen 23:00 uur en 10:00 uur) Ja / Nee 10. Heb je last van slaapstoornissen zoals?:
REM slaap gedragstoornis Ja / Nee Insomnia
Narcolespie Parasomnia
Andere slaap stoornissen
10. Gebruik je medicatie? Ja / Nee
