The Real-World What-Where-When memory test is a novel episodic memory test, in which participants need to recall which objects have been hidden in which locations on which of two distinct occasions. It is easy to run and is sensitive to normal cognitive aging.
Episodic memory is a complex memory system which allows recall and mental re-experience of previous episodes from one's own life. Real-life episodic memories are about events in their spatiotemporal context and are typically visuospatial, rather than verbal. Yet often, tests of episodic memory use verbal material to be recalled (word lists, stories). The Real-World What-Where-When memory test requires participants to hide a total of 16 different objects in 16 different locations over two temporal occasions, 2 h apart. Another two hours later, they are then asked to recall which objects (What) they had hidden in which locations (Where) and on which of the two occasions (When). In addition to counting the number of correctly recalled complete what-where-when combinations, this task can also be used to test real-world spatial memory and object memory. This task is sensitive to normal cognitive aging, and correlates well with performance on other episodic memory tasks, while at the same time providing more ecological validity and being cheap and easy to run.
Episodic memory is memory for unique events from one's own past that are experienced as a reliving of the original event (mental time travel)1,2. It is also one of the first types of memory to be affected in the early stages of many forms of dementia3,4. The medial temporal lobe, and more specifically the hippocampus, is believed to be an important structure in the processing of episodic memories5, and therefore any conditions that affect hippocampal function, like aging and many mood disorders, are also believed to affect episodic memory function. As such, episodic memory function can be a useful biomarker for a range of neurological and psychiatric conditions6.
Methods for quantifying episodic memory, however, are still less than ideal. Real-world everyday episodic memories are integrative memories of unique events in their spatiotemporal context7, usually incidentally encoded4. The two most common methods used in both the clinic and in academic research are word list learning8 and recounting a story from one's own past3. Both methods have advantages and disadvantages. The advantage of the word lists over the story approach is that the assessor knows exactly what the right answers are. This is difficult to assess with spontaneous stories from the participant/patient's past, since often no objective evidence is available and even accounts from family members may have incorrect details in them. The advantage of the stories is that they actually assess the typical content and structure of episodic memories: events in spatiotemporal context, with information about What happened, Where and When bound together7. Word lists do not require any context to be recalled at all, and are often rehearsed several times (e.g. the Rey Auditory Verbal Learning Task).
Recently, several attempts have been made to construct episodic memory tasks that combine the strengths of the two classical tests while minimizing the drawbacks9,10,11,12,13,14,15,16,17,18,19,20,21,22. The current protocol is the most recent version of a What-Where-When episodic memory test which has been developed at Newcastle University10,16,22. The concept is based on the work with non-human animals, started by Clayton and Dickinson23, and adapted for work with a range of other species24,25,26,27,28, some of which have confirmed the sensitivity of this paradigm to medial temporal lobe damage29. It is one of several attempts at incorporating a What-Where-When framework in episodic memory testing with adult humans18,20,30,31, but the only one to be performed in a real environment, without the use of computers, making it easy for participants/patients to engage with and low-cost to carry out.
This protocol was approved by Newcastle University's Faculty of Medical Sciences Ethics Committee (approval number 515_1).
1. Preparation for the study
Figure 1: Object identity sheets. These are the sheets that are placed next to the pile of 20 objects. The left-hand sheet is placed there in phase 1 and the right-hand sheet in phase 2. Participants are supposed to pick up the objects from the pile in the order indicated on the sheet. Note that in phase 2, all the objects from phase 1 have been collected up and added to the pile again, so that during phase 2, participants again search through a pile of 20 objects to start with. Please click here to view a larger version of this figure.
2. Session 1
3. First break
4. Session 2
5. Second break
6. Session 3
Figure 2: Subjective experience reporting. After reporting on their memory for objects, locations and phases, the participants are asked to complete these two scales, which report on the subjective experience of their memories (top scale) and on how much they actually rehearsed the material after each hiding phase. Please click here to view a larger version of this figure.
7. Data extraction and analysis
Older people (65+) remember fewer complete What-Where-When combinations than do younger people (18-25; Χ2(1) = 9.5; p = 0.002; Figure 3). Note that although as a group, older people perform worse than younger people, there are some older people who perform as well or better than some young people. This variation may be informative if it is predictive of other conditions.
One can also investigate how other episodic memory tests predict performance on the memory for What-Where-When combinations. For these representative results, the results of Kessels' Object-Location task32 are presented. This task has several components, including Combined Object Memory (COM), in which 10 different objects have to be remembered and replaced in their exact positions on an otherwise empty computer screen. In this version, the participants had 3 minutes between studying the layout of the 10 objects, and the test in which they had to recreate this layout. During these three minutes, they performed another task, in order to prevent them from holding the information in working memory. Individuals' performance on the COM task significantly predicted the number of correctly recalled WWW combinations (Χ2(1) = 6.27; p = 0.012). The slope of the regression line is steeper for older people than for younger people (Χ2(1) = 4.97; p = 0.026; Figure 4).
Figure 3: Age differences in WWW memory. The figure represents the total number of correct What-Where-When combinations remembered (out of 16 possible combinations) by all participants in the two age groups. The size of the symbols represents the number of individuals that remembered that number of combinations. Please click here to view a larger version of this figure.
Figure 4: Prediction of WWW memory by other memory tasks. The number of correct What-Where-When combinations remembered by each individual is significantly predicted by the accuracy with which these individuals can remember and reconstruct a spatial array of 10 objects on a computer screen (Kessels' Combined Object Memory or COM task). The performance measure for the COM task is an error score which indicates how far away (in mm) the objects were placed from their correct locations. Errors for all 10 objects have been summed to obtain one error score per participant. Please click here to view a larger version of this figure.
The data show that performance on other tasks which are supposed to measure episodic memory predicts performance on the Real-World What-Where-When memory task as well. However, these correlations are likely to represent a shared subset of cognitive abilities used by the different tasks. The Real-World What-Where-When memory task has the advantage over these other tasks in that it tests people's memory for two actual events which happened in a real spatio-temporal context. Unlike asking people about events from their own lives, however, in this case, the experimenter or clinician knows exactly what happened in the event, since they set it up. This gives the task an ecological validity not shared with most other tasks, as even those that also apply a what-where-when framework, typically do this on a computer, thereby losing the immersion aspect of a real-world experience18,20,30,31. Only immersive virtual reality might combine the advantages of computer stimulation and a real-world-like experience, but this type of equipment is not readily available yet to most people, while the Real-World What-Where-When memory task is easy and cheap to run. For even more ecological validity, the task can be run as an incidental encoding task, which is impossible to do when asking people to memorize lists of words or word pairs. However, it can only be used to test incidental coding once. After the participants have taken part once, they will forever know that this is a memory task.
The task also has the advantage that it has several different outcome measures all from one experience: it can test purely spatial memory and purely object memory, as well as memory for the binding of the different elements. These different aspects could potentially be useful to separate different neuropsychological problems in patients. The task has drawbacks as well. The three phases have been compressed within a day, but to test long-term memory, the whole task takes at least 4.5 h, usually a bit longer. This is not a problem when people are in the clinic or the testing environment for at least half a day anyway, but does constrain the circumstances under which it can be used. Single hiding episodes (essentially ignoring the temporal component of episodic memories) or shorter retention intervals are easy enough to implement and it would be interesting to find out how they affect performance on the task. Another drawback of having a manual task in a real-world environment is that scoring the outcome by hand takes longer than computerized tasks that can give a performance score immediately. That being said, counting up the number of correctly remembered what-where-when combinations takes very little time. It is scoring incomplete combinations that may take a bit longer.
The task as presented here can be used to look at other aspects of episodic memory as well. One could analyze the memory for the sequence in which the objects were hidden, and/or the effect of memory decay and interference between the two phases on the memory traces. Indeed, by adding more phases to the experiment, this simple memory task could be used to investigate quite detailed hypotheses about the temporal aspects of memory, including for example the scale invariance of the memory decay parameters15.
In summary, this simple cognitive test of episodic memory has more ecological validity than existing tests, yet seems to distinguish similar groups as have been shown to be impaired in episodic memory using other tests. Further studies are needed to understand how sensitive this task is and whether it might be useful in, for example, early diagnosis of cognitive impairments.
The authors have nothing to disclose.
Thank you to all the participants that have helped develop this methodology over different iterations of the task. Thank you also to all the students who helped by running the different iterations over the years: Natasha Dubes, Emma Denning, Victoria Bellhouse, Stephen Holland, Melissa Anderson, Katie Shaw, Sarah Morgan, Karla Butterworth, Michael Craig, Lauren Wray, Olivia Sanderson, Daniel Lai, Rajameenakshi Boopathy and Chun Kit Ho. This research was funded by Newcastle University’s contributions to student research projects.
These materials are just examples. Any 20 easy-to-name small objects will do. |
– A tea light |
– A toy digger |
– A toy frog |
– A clothing hook |
– A spoon |
– A set of keys |
– A button |
– A lip balm |
– A toy snowman |
– A bottle cap |
– A die |
– A lighter |
– A comb |
– A wrist band |
– A padlock |
– A butterfly pin |
– A ruler |
– A guitar plectrum |
– A battery |
– A USB stick |