Is human memory reliable?

Here’s a memory fact for you: the human memory is NOT like a video camera.

This would imply that the visual information we receive in our eyes is encoded and stored, without any further processing, to just be recalled in exactly the same way we first perceived it. This is not the case.

One theory states that our long term memory is Reconstructive. This means that abstract principals about the input material are stored and the memory is then reconstructed according to these principals during recall. An experiment to show this was done way back in 1932 by Bartlett, who showed a bias in picture recall to real life objects, when the pictures were actually abstract. Here is an example of some of the stimuli and recall images used in his experiment:

As you can see, participants’ memory of the original object was skewed towards whichever real life object they thought it looked like. This are simple shapes, but yet the participants were unable to recall them accurately.

The interference theory provides strong evidence that human memory isn’t reliable as it shows that our memories can be altered by previous learning (proactive interference) or by new learning (retroactive interference). There are several examples of this is real life: for example, when you get a new phone you find it difficult to type as well as you did on your old one (proactive), and if later for some reason you try to type on your old phone again you will also find that difficult, as you are used to the new one (retroactive).

Isurin & McDonald (2001) carried out an experiment using bilingual participants to assess the effects of retroactive interference. They presented participants with a picture and corresponding word in either Russian or Hebrew, and then gave half the participants the same pictures but with the words in another language. Recall was then tested, and it was found that it became worse the more learning trials they had using words in the 2nd language. They concluded that these results showed retroactive interference, and hypothesised that this could be why many bilingual individuals forget their first language if it is not used.

More evidence that human memory is not reliable come from the fact that you can have ‘false memories’ – memories you are certain are true but did not actually happen. One of the most famous studies to show this was carried out by Roediger & McDermott (1995). They gave participants lists of words with something in common and found that 40% of the time participants recalled this theme as one of the words, even if it was not presented in the list. For example: participants were given sky, wet, cloud, puddle etc, and recalled the word rain. When they asked participants why they did this after the experiment, nearly 3/4 of them said that they had a strong memory of the word being in the original list (not just a feeling that they had seen it somewhere). This shows how easily memory can be manipulated.

While these may seem like pretty trivial examples of incorrect memories, this topic has some really important real-life applications, such as eye-witness testimonies to a crime. Elizabeth Loftus is a leader in this field, and had done many experiments to show the optimum conditions for eyewitness recall. Misleading questions are a type of retroactive interference which can alter an eyewitness’ memory.

In her well known experiment, she found participants were more likely to overestimate speed if they were asked ‘how fast were the cars moving when they smashed into each other?’ when referring to a video, rather than if the word ‘smashed’ was replaced with ‘bumped’. This shows that interviewers must be extremely careful about the questions they ask eyewitnesses.

Another experiment by Loftus illustrated ‘weapon focus’: this is the impairment of memory caused by the eyewitness focusing on a weapon and less on other details. they asked participants to watch one of two videos, one containing a gun and one without. They found that participants had worse memory of the scene in the video containing the gun, suggesting that weapon focus had occurred.

Unfortunately, there is also evidence that racial/social stereotypes can also affect a persons’ memory of an event. Lindholm & Christianson (1998) found that Swedish students were more likely to identify an immigrant, rather than a Swede from a line up when asked to identify the person carried out a crime in a video they were shown.

So, the question is: do you trust your memory?



Memories: everyone has them, and probably takes them for granted. You think that you can always rely on them, although it does let you down occasionally (ever walked into a room and forgot why you went in there?!). So how reliable actually is it? Here’s my explanation of human memory.

You can separate memory into two different types: short term memory and long term memory. In this post, I’m going to focus on long term memory, just because I think it’s more interesting.

Here is a diagram explaining how human long term memory can be broken down into various different subsystems:

As you can see – there is more to long term memory than you’d expect. The two types of long term memory are declarative (memories you can vocalise) and non-declarative (memories you’d fine hard to talk about or explain). Examples of declarative memories include semantic memories (e.g. facts) and episodic memories (personal memories). These differ from non-declarative memories such as procedural memories (e.g. how to ride a bike) or emotional memories (e.g. feeling happy). These sub-systems of memory help researchers to classify memory deficits and measure how they are learnt.

Hopefully, you now know a bit more about the things which make up our long term memory. However, there is more to its different components than the memories they contain – these different types of memory are stored in different areas of the brain. Although the range of locations in the brain used to store memories can get quite confusing, I’ll aim to make it as simple as possible.

The hippocampus is a structure in the middle of the temporal lobe on both sides of the brain, and is vital for the storage and recollection of episodic and some semantic memories. This is known because of research on case studies, such as patient H.M. whose hippocampus was removed to cure his violent epilepsy. He was left unable to form new declarative memories (anterograde amnesia) and his declarative memories from the years just before his operation had also disappeared (retrograde amnesia). However, he was still able to carry out motor functions such as walking, in other words, his procedural memory was still intact. This therefore shows that procedural memories must be stored in another area of the brain.

As well as the hippocampus, it is also believed that the structures around it such as the mammillary bodies, fornix and perirhinal cortex are also important for remembering. The perirhinal cortex is thought to be especially important for semantic memory. The frontal lobe is also thought to play a part in memory, especially for memories which have been rehearsed several times, and perhaps date back years. It is thought that stronger traces are developed as each memory is recalled over and over, which means that it becomes less dependent on the hippocampus.

As for procedural memory, this is thought to be reliant on several structures in the brain that are involved in motor-learning, such as the basal ganglia and the cerebellum (the structure at the back of the brain). The amygdala are next to the hippocampus and are involved in processing emotional memory, something which is also done by the frontal cortex.

I hope you now understand human long term memory a bit better than you did before you read this post! As you can see, it relies upon many structures in the brain: there is no one area responsible for our memories. Check back for future posts about memory impairments, and how reliable our memory actually is 🙂