LECTURE 6-History of memory research and sensory memory
1-What is memory?
Memory is the mental process or ability to encode, store and retrieve information.
From this definition, we can see that memory involves three stages:
Encoding stage: you get information into your memory system.
Storage stage: you retain information in your memory.
Retrieval stage: you get information out of memory. When you need to recall something, you need to use your
existing knowledge to learn something new.
So, memory is essential to our lives: without it we could not recognise others, talk, read, or write because we would
remember nothing about language, for instance. We would have limited personalities too, since we would not have memory
of our experiences.
2-Memory research
2.1-Hermann Ebbinghaus (1850-1909)
He was the 1st famous psychology that pioneered the experimental study of memory. He was the first one that scientifically
studied memory since before memory research was mostly conducted by philosophers and was based on observational
descriptions and speculations.
He published his monumental book Memory A Contribution to Experimental Psychology (Über das Gedächtmis in German)
in 1885, one of the most influential psychology research on memory.
He created these so-called nonsense syllables to study memory, it is, syllables that always consist of a consonant, a vowel and
a consonant but that make no sense and have no meaning associated. He did that not to get contaminated results with bias
since any participant has prior knowledge of the words.
2.2-Carl Lashley (1950)
He is an American psychologist that first tried to locate the area in the brain where memory trace was stored. He tried to find
what he called engram, a single brain region where memory is stored.
He trained the rats to ran through a maze and then he systematically disrupt different parts of the rat's brain. By cutting the
rat's brain in different places in order to see after which part of the brain was removed, the rat lost the ability to run through
the maze. After a long series of studies, he found basically that lesion a specific area of the rat's brain did not affect the maze
running performance at all, but what really matters is the amount of the cortex that has been removed. So, the more brain
tissues being removed, the more memory storage or the more performance was damaged.
This graph shows the percentage of the cortex removed from 10% to 70% and
the number of errors the rat made in maze running. One has to really remove
quite a lot of brain tissues before the rat start to make a lot of errors.
So, he conclude that memories are not localised in any specific brain regions,
but they are widely distributed across the cortex and developed the idea of
equipotentiality, in which he concluded that all parts of the cortex contribute
equally to complex behaviours like learning and memory.
What we know today is that Lashley's conclusion was both true and untrue:
memory, from what we known today, is indeed widely distributed, but he
wrongly assumed that memory was unitary (there is only one memory and
one type of memory trace stored all over the brain).
The problem of Lashley's study is that potentially there could be many brain circuits involved in a complex act like running a
maze and eliminating one part of the brain is not enough to disrupt the entire act. So, for example, different parts of the maze
may look differently, feel different, smell different, or perhaps even sound different. If some of a part of the maze was noisier
than others and the rats could rely on all these different types of memory cues, when you just remove one part of the brain,
there's only one trace or one cue disrupted. The rat could still rely on other cues to run through the maze.
,The current view in memory research is that memory is distributed in many areas in the brain, but it's not equally distributed
in all areas. There are some selective memories in some specific parts of the brain, but mostly the brain is working in concert
to remember certain events.
What processes a memory of an event rarely just involve one type of memory, it's a systematic coordination across all
different memory systems. But where exactly different memories are stored is still unclear so far and there's still ongoing
debates in neuroscience.
2.3-Donald Hebb (1904 – 1985)
He's a Canadian psychologist, one of the forerunner of the computational neuroscience who tried to mathematically model
our brain activities and he's well-known for his theory on the cellular basis of learning and memory.
According to him, electrical activities of neurones in our brain, which is associated with all the sensory memory-motor
cognitive experience, will leave an imprint in the neuronal or synaptic structure and modify these structures. This
modification will persist for a long period of time getting the form of learning and the memory.
His book entitled The Organisation of Behaviour has become a classic textbook in the field of neurobiology. He proposed
that, if a neuron repeatedly becomes activate at about the same time that another neurone fires, the connection between these
two neurones will be strengthened. This process is called synaptic plasticity. So, he was trying to claim that cells that fire
together will wire together.
Unfortunately, at his time, he couldn't test his hypothesis because the technology was not available at that time.
2.4-Terje Lømo (1966)
The breakthrough came from Terri Lomo, who discovered a very striking
phenomenon called long term potentiation, a process thought to be the
cellular and molecular underpinnings of memory.
This important discovery revolutionised our thinking of how learning and
memory could occur in the brain. It directly fits the line of Hebb's
postulates, so a great deal of functions of memory happen in the change
of one neuron to communicate with another neuron.
When the two neurones are activated at the same time, they kind of
connect to each other and this is how you learn information and store
information in your memory.
Long term potentiation has been found in many areas associated with learning and memory, including hippocampus and
amygdala.
Now scientists have created mutant or genetically modified mice so that long term potentiation in their hippocampus does not
work well and found that these mice have shown worse memory ability than other mice.
Demonstration of LTP and Neuroplasticity: https://www.youtube.com/watch?v=8Vo-rcVMgbI&spfreload=5 (from 1.10min)
2.5-Modern memory research
The multi-store model (Atkinson & Shiffrin, 1971):
One of the earliest important model for memory. There are other type of memory models, this is just one of the influential.
The 1st stage of the model is getting the input
in sensory registers or sensory stores. This is
where you first interact with the information,
your environment. It's just a very temporary
register of all the information, you sense your
senses taking in.
We have a lot of information coming into our
sensory memory, but we can't process all of
them. We have to decide to only pay attention to certain information and then that information will be passed on into short
term memory store. Short term memory works a bit differently from sensory memory: it is less restricted by time, but more
restricted by quantity (± 7 pieces of information at a time).
, And then if you further process it, it will move out to long term memory store. Once information gets in there, it's like the
save button on your computer. Our memory does not work always perfectly, but lots of information can be stored in the long-
term memory.
3-Sensory memory
Sensory memory allows us to hold information from different sensory modalities (visual, auditory, olfactory, tactile and
gustatory) for a very brief period of time.
The information is held just long enough for us to decide, decide whether to attend to the information and pass it on to short-
term memory and if not, the information will decay rapidly.
Even though we have five sensors, the two most studied modalities are visual and auditory modalities. Within sensory
memory, we have iconic memory of all the visual sensory memory and the echoic memory for the auditory sensory memory.
3.1-Iconic memory
George Sperling (1960):
To study the capacity of iconic memory, George Sperling (1960) presented participants letter array (3 rows, 4 letters/row) for
a short-period of time (.015 - .5 sec) and asked them to report all the items they saw. What has been found in that study was
that participants on average could report about 4.5 letters and because there were in total 12 letters, participants could
remember around 37.5% of the information.
However, there is a problem here because information is lost so quickly that maybe subjects couldn't report it fast enough
before it decayed. So how do we know how much information is available and how long it lasts? This will require the partial
report technique: Sperling presented participants arrays of letters and once they disappeared, asked them to report just a
subset of letters, like first or second or third row depending on whether they heard a high tone or a medium tone or a low
tone.
So, instead of report the letters from the very first one and while you reporting it, your sensory memory loss information
while you're recalling, it didn't ask you to start from the beginning, but it will give you a cue. This cue (high medium or low
tone) will tell you whether you need to recall the letters from the first row or the second row or the third row.
If participants indeed could only remember like four letters, they have no chance to recall anything from the third row or the
second row. However, if they could encode or remember more information in their sensory memory, they should be able to
report the second rows letters in the second row or even in the third row.
What was found in the original study was that participants could, on average report 3 out of 4 letters in each trial regardless
of whether it's the first or second or the third row. This equals actually 75% of all the letters, meaning that means participant
must have, on average, about 9 out of 12 letters available in their iconic memory.
The second question is, how long does iconic sensory memory last? In order to study how long iconic memory last, Sperling
in his study varied the time between the disappearance of the displayed letters and the onset of the cue tone.
In a trial, participant was first presented with the 12 letters for only 50 milliseconds, and then the time interval between the
disappearance of these letters and the onset of the cue tones varied between zero second to one second. The results showed
that initially, if the cue tone was played immediately, people with partial recall technique could report around nine letters
correct (much higher than the whole report method) and then with longer interval the partial report performance decreased. In
about 1 second, the partial report performance decreased to the level that corresponds to the whole report procedure.
Therefore, it is concluded that iconic memory can last about one second.
3.2-Echoic memory
Darwin, Turvey, and Crowder (1972) designed a study that's similar to the
iconic memory study: presented simultaneously a different list of three
consecutive items from each of the three different spatial locations (left ear, right ear and then middle) and then either
