I would be likely to recall the last fruit I had but it could be so that I recall better the fruit I eat more often or daily.  Test the two conditions. Here is an online test:

https://www.psychologistworld.com/memory/test1.php

It depends what you mean familiarity. There is a large amount of evidence that the frequency that a person encounters a word has a significant impact on episodic memory in younger and older adults. You can find some research on my publications.

The major problem is that it's difficult to discriminate between familiarity and frequency, especially when you come to words. Also, how would you get an objective measure of familiarity? For example, in the USA you call an aubergine an eggplant but in the UK we will not be familiar with the word eggplant but we would be more familiar with the word aubergine and vice versa. It is also difficult to measure familiarity between cultures, ages of participants and regional areas.

The other difficult part of your question is how do you define retrieval? In the context that you phase the question then it is suggesting cued-recall, but you would also want to investigate uncued recall and maybe recognition. One in interesting finding has shown that low-frequency words are easier to recognise than high-frequency words, but with recall the opposite is true. That could be worth thinking about.

Hope this is helpful

Nick

Word-frequency effects dominate most tasks, but then, most memory tasks in lab settings require you to recall an already presented list of words, or a word-pair that you have already encountered. So, I think episodic memory may not be an ideal domain to explore word-frequency effects.

By contrast, semantic memory is a better domain to explore these. For example, in tasks like category fluency where participants have to recall names of as many animals as they can in one minute, healthy people usually recall animals like 'dog', 'cat' etc., first and then may move on to less frequently encountered animals like hippo, iguana etc (unless they are implicitly using a certain strategy like moving alphabetically). In the pattern of responses they provide within a minute, you are likely to see more high-frequency words (animals they are more likely to come across) than low-frequency ones. A nice lesion model to observe this is semantic dementia (SD), where patients start losing conceptual knowledge for lower frequency words with disease progression. If you were to do a category fluency task on these patients, they will more likely recall animals like dog, cat as with the breakdown in their semantic system, words like hippo and iguana do not hold any conceptual meaning anymore. 

Your example of recalling a fruit I have eaten, is a close interaction between semantic and episodic memory - which though support each other, are also very different from each other. While I may recall the experience of eating a fruit (episodic), the knowledge about the fruit is semantic. An SD patient would be a classic fit for testing this breakdown, as even if they ate an eggplant, they might tell you they ate a 'long, purple thing' (grasping for words to describe the eggplant), or in more moderate-to-severe cases, might just tell you they ate a banana (as the conceptual knowledge for that may be preserved over an eggplant) as its more typical and frequently encountered.

Hope this helps!

Hi Siddharth,

I understand your reasoning but you are actually testing verbal fluency and not recollection. Verbal fluency relies on the PFC and does not really tell you much about recollection but more about executive functioning.

One way round this would be to ask healthy individuals to eat various familiar and unfamiliar items one week and then invite them back next to list which items they have tried and if you wanted to use a metacognition task you could ask them for how confident they were on a scale of 1 to 7 and then conduct gamma correlations.

If this provided any significant findings then you could expand it to older adults and participants which who discussed. I think this would tap into recollection much more than a verbal fluency task?

I think some would argue that categorization is memory - or at the very least, a process of memory.

I believe the *Episodic Simulation Hypothesis (http://scholar.harvard.edu/files/schacterlab/files/schacter_addis_buckner_anyas_2008.pdf) would point whatever fruit being more salient to the individual as what they will remember. So, in a way, this would be the more typical fruit. So typicality can be, in a way, relative to one's own categorization of exemplars and prototypes (expertise).

Maybe this Scholarpedia page has some answers for you: http://www.scholarpedia.org/article/Adaptive_resonance_theory

* The Episodic Simulation Hypothesis is focused on the mental simulation of behavior in the future, but evidence points to this ability being majorly dependent on one's ability to remember carrying out that action in the past.

Hi Siddharth,

I have actually worked with Martin Conway when he was at Leeds and many of our conversations were about how one can't discriminate between episodic, autobiographical and semantic memory. It is very difficult to separate the three types of memory, especially episodic and semantic memory. For example, I cannot stand bananas therefore if I did a verbal fluency task on fruits then bananas would be one of the first fruits that I would name. On the other hand, I like citrus fruits so I would also name a lot of citrus fruits. It is difficult to assess whether this is due to episodic, autobiographical, or semantic memory. It is possible that some individuals would just say "oranges" but there are very different types of oranges.

Second, there could be a way of combining both techniques. One could do the verbal fluency task first and then do the experimental task two or three weeks later. The experimental task based the episodic memory task on the verbal fluency task so participants are only tested on the fruits that they have come up with during the verbal fluency task... perhaps you could also add in some fruits that they did not come with during the verbal fluency task and see if these get a higher JOL in the subsquent recall task.

I agree with you that it depends on what the participants might have eaten that week, but you could also do post-hoc analysis and check the grocery receipts and use this as a covariate in post-hoc statistical analysis?

Hi Nicholas,

Based on the information that you mentioned, I would say that is more likely to recall a typical exemplar, for instance apple, instead of an atypical one, such as a fig.

My argument is grounded on the idea that the task, recalling a fruit that you have eaten, will direct participants to their representation of a fruit. Since an apple has bigger feature overlap (has a peel, is sweet and so on) with the common idea of what a fruit is, then participants may search their memory for typical exemplars, awhile atypical items may be recalled less well because they are not as closely integrated with the category because they share fewer features (doesn´t has a lump).

In the literature you may find opposite results for what kind of exemplar is remembered better (typical or atypical). In my opinion, these contrastive results may be due to the episodic task required. For instance, Schimdt (1996) and Whitney et al. (1983) found better memory for typical exemplar in a free recall task (and I would say that your task is similar to a free recall task). On the other hand, others have found no typicality effect (e.g. Greenberg and Bjorklund, 1981). Awhile I have found the reverse of the typicality effect, better memory for atypical than typical.

Another study, from Whitney et al. (1983) showed greater clustering for typical than atypical items, which seems to support the idea that typical items are more closely integrated in the central tendency (feature overlap) of the category. In contrast, during recognition judgments, item distinctiveness, as promoted by idiosyncratic features of the atypical items (meaning that they have other features that do not fit so well with the category, and because that they are considered atypical exemplar) may be the key to enhance retrieval.

Another thing is to consider the specific experience of the participants. Because if you are expecting that they actually recall what fruit they have eaten you will find t a problem regarding to the variety and frequency of fruit consumption for each participant.

Maybe you also find interresting the following study from may lab: Garcia-Marques, L., Nunes, L., Marques, P., Carneiro, P., & Weinstein, Y. (2015). Adapting to test structure: letting testing teach what to learn. Memory, 23, 365-380.

wherein were found different results for typicality depending on what were asked to do on the encoding task previous to the episodic retrieval.

HI Mara,

I agree with you regarding recall and recognition but I am just finding it difficult how one would seperate typicality and frequency? If we are talking about fruits then I would imagine typicality would be quite personal, but when we are dealing with frequency then there are ways to come up with objective measures of frequency.

Hi Nicholas Almond,

Just to clarify, my previous comments were an attempt to answer Nicholas Coles’ question.

Regarding how to disentangle typicality and frequency, one possible way is to get measures for typicality (ask for a set of exemplars how typical they are for their specific category) and then use normative data for word frequency. The idea would be to have both typical and atypical items varying orthogonally with frequency (although the majority of typical exemplars tent to have high familiarity and also high frequency ratings). But looking at normative data in Portuguese, both orange and grapefruit are considered typical fruits (although orange has a higher value then grapefruit), however, in Portuguese, orange is much more frequent then grapefruit. So, at the end, it would be possible to have two groups of items, one with typical and other with atypical items, that do not differ in terms of frequency.  

In my master research, I faced this problem. One year later, I still had the same question, I have created two different groups that did not differ in frequency, but each item had a specific typicality value and also a frequency value. Could or not frequency still influence my results? The way that I found to answer my question was to do a regression analysis using typicality and frequency as IV. And long history short, I only found a significant influence of frequency, along with typicality, for RT for one of my episodic conditions. So, with the materials that I have used, frequency did not affect ACC or RT in the semantic/encoding task. It only affected RT in one of the two episodic/retrieval conditions, along with typicality, meaning that participants take more time to respond to more atypical and more infrequent items, but this not seems to affect how well they remember the item (well, if we reduce the time to answer, maybe the results will be different).

This answers your question? I am not sure if I grasp the previous discussion on the frequency topic. I only get excited to answer because typicality was the theme of my master thesis, I am more than happy to talk about it.

Hi Mara,

Thanks for clearing that up