Memory outside the hippocampus

ScienceDaily reports (here) on a paper by Bindschaedler, Peter-Favre, Maeder, Hirsbrunner, Clarke in Cortex, Growing up with bilateral hippocampal atrophy: From childhood to teenage.

A child with damage to the hippocampus at birth did have some memory: memory of events was affected but not memory of general knowledge. His episodic memory was almost 90% gone while his semantic memory was within normal levels. This means that the two types of declarative memory (memory that can be accessed consciously) are different and that semantic memory is not just a consolidation of episodic memory.

The paper also seems to discuss evidence that recall and familiarity are separate.

Here is the abstract of the paper:

The respective roles of the medial temporal lobe (MTL) structures in memory are controversial. Some authors put forward a modular account according to which episodic memory and recollection-based processes are crucially dependent on the hippocampal formation whereas semantic acquisition and familiarity-based processes rely on the adjacent parahippocampal gyri. Others defend a unitary view.

We report the case of VJ, a boy with developmental amnesia of most likely perinatal onset diagnosed at the age of 8. Magnetic resonance imaging (MRI), including quantitative volumetric measurements of the hippocampal formation and of the entorhinal, perirhinal, and temporopolar cortices, showed severe, bilateral atrophy of the hippocampal formation, fornix and mammillary bodies; by contrast, the perirhinal cortex was within normal range and the entorhinal and temporopolar cortex remained within two standard deviations (SDs) from controls’ mean. We examined the development of his semantic knowledge from childhood to teenage as well as his recognition and cued recall memory abilities. On tasks tapping semantic memory, VJ increased his raw scores across years at the same rate as children from large standardisation samples, except for one task; he achieved average performance, consistent with his socio-educational background. He performed within normal range on 74% of recognition tests and achieved average to above average scores on 42% of them despite very severe impairment on 82% of episodic recall tasks. Both faces and landscapes-scenes gave rise to above average scores when tested with coloured stimuli. Cued recall, although impaired, was largely superior to free recall.

This case supports a modular account of the MTL with episodic, but not semantic memory depending on the hippocampal formation. Furthermore, the overall pattern of findings is consistent with evidence from both brain-damaged and neuroimaging studies indicating that recollection requires intact hippocampal formation and familiarity relies, at least partly, on the adjacent temporal lobe cortex.

Another look at LIDA

One of the interesting things about the Madl, Baars, Franklin LIDA model is the number of memory stores that it envisages. I have thought of consciousness as the ‘leading edge of memory’, at least of episodic memory. Hence my interest in the model’s use of memory.

Let us walk through their cognitive cycle to see what forms of memory are mentioned.

In the perception part of the cycle, the percept is held in the preconscious buffers of LIDAs working memory (workspace) and temporary structures are built. I assume the temporary structures are a preliminary world model. This seems definitely wider than the local sensory memories like visual memory.

The residual contents of the working memory is associated with the incoming percept and with associations from episodic and declarative memory in order to create an updated working memory. In summary, a number of memory stores are associated to give the current workspace: the new percept, the previous workspace, episodic memory, declarative memory. But they do not include and input from a motor system memory.

This resulting workspace is examined by attention processes to bring some parts of the workspace into consciousness. These parts (the novel, relevant, urgent and insistent) are transferred to another memory, the global workspace. This is equivalent to conscious broadcast as the contents of the global workspace are available to many process in the brain.

A procedural memory is postulated, hold various relevant behavior schemes which can retrieve the information they require from the global workspace. Goals are adjusted, actions selected and taken. Results of the action are fed back into the cycle through effects on the environment being included in sensory input.

In summary and in different words we have: individual stores in each sensory mode, a store with the preliminary precept, a store for finished world model, stores for episodic/declarative long-term memory, a store for ongoing behavior plans/goals, a store of action implementation procedures. This model gives a look at the sort of memory stores that are likely to be needed for cognition.

I have a big reservation about this cycle (although in general it is attractive), it does not deal well with prediction. For improvement, if we start with an action procedure rather than incoming sensory information and draw the cycle from that starting point. The action procedure, as well as, and before, producing the action, also is integrated into the existing workspace so as to create a model of the world as it is expected in a fraction of a second. What fraction of a second? Long enough for the conscious broadcast of the global workspace to coincide with the prediction - so that ‘now’ is ‘now’. The incoming sensory information can be compared with the workspace to register any errors and allow the action procedure to be amended if required. In other words, sensory and motor systems have two links: there is an external link through the environment and an internal one through prediction and they are compared in the workspace.

Madl, T., Baars, B., & Franklin, S. (2011). The Timing of the Cognitive Cycle PLoS ONE, 6 (4) DOI: 10.1371/journal.pone.0014803

What change blindness says about memory

In change blindness some part of a scene is changed and the change is not noticed by the observer. This can happen when the change is not happening on the retina in a stable condition. It can happen when there is a mask (a blank screen to fast to see), a blink, an eye movement, a change in point of view, an interruption in a action and so on, anything disrupts the continuity of the retina image.

There was some questions of whether change blindness could happen with objects that were at the center of attention. Surely, if you were engaged in a conversation with someone, they could not be replaced with another person without the change being noticed. But they can as Simons and Levin showed in their paper (see citation).

What does this say about our memories? Simon and Levin say:

If we constantly noticed such changes, they would likely detract from our ability to focus on other, more important aspects of our visual world. Change detection as a method relies on the tendency of our visual system to assume an unchanging world. The fact that we do not expect one person to be replaced by another during an interaction may contribute to our inability to detect such changes. … Taken together, these experiments show that even substantial changes to the objects with which we are directly interacting will often go unnoticed. Our visual system does not automatically compare the features of a visual scene from one instant to the next in order to form a continuous representation; we do not form a detailed visual representation of our world. Instead, our abstract expectations about a situation allow us to focus on a small subset of the available information that we can use to check for consistency from one instant to the next.

In effect we delude ourselves as to the completeness of our immediate memory. We remember the unexpected if we notice but there is no guarantee that we will notice.

Simons D.J., & Levin D.T. (1998). Failure to detect changes to people during a real-world interaction Psychonomic Bulletin and Review, 5, 644-649

Memory as a resource

Some people are surprised, even disturbed, by the idea that our vision does not give us an accurate picture of what we look at. For example, the colours we experience are not a measure of the wavelength of the light entering our eyes. But accuracy is not the point of vision; the function is to be useful and colour consistency is far more useful then fidelity to wavelength spectra. The same surprise is shown in the reaction to the idea that our memories are reworked continuously so that over time they lose their accuracy. This is not a fault in memory. Again the reason we store memories is to have a useful resource, not necessarily one with detailed accuracy. A great deal of biological energy is used to create memories and to re-consolidate them and therefore we can assume that they have a very important biological role.

In order to have episodic memory, we first have to have the experiences to remember – we create consciousness experiences which become stored as temporary memories. These are soon made more permanent as an event or an episode. Over weeks, months, years, decades they are continually reworked and re-consolidated. They are packaged together and lose their individuality, they are up-dated by newer memories, they are categorized and lose detail not related to their category. Eventually they lose their character as episodic memory and became more a factual or semantic type of memory. What has happened in all this change is that we have learned from experience. This in itself would probably justify the biological cost of consciousness and memory but more has been proposed by Moshe Bar, Donna Rose Addis, Daniel L. Schacter and others. They have put memory at the center not just of the past and learning but of the present and future, of prediction, understanding and cognition.

Bar envisages a ‘proactive brain’ which builds analogues by examining what something is like. If A is like B then they share an analogy which grows as other like things are remembered. Each memory added to a particular analogy brings associations with it, so the associations of each analogy grows. The associations of an analogy are in effect predictions of what else will be found along with the analogy. Thus memory is the material of prediction and foresight. Between memories, predictions and idle imaginings we have the simulations we need to plan our actions. These simulations can even become ‘scripts’ for guiding behaviour and sets of scripts to determine a ‘mindset’ appropriate for a particular type of situation.

Schacter and Addis call their scheme the ‘prospective brain’.

A rapidly growing number of recent studies show the imagining the future depends on much of the same neural machinery that is needed for remembering the past. These findings have led to the concept of the prospective brain; an idea that a critical function of the brain is to use stored information to imagine, simulate and predict possible future events. We suggest that processes such as memory can be productively re-conceptualized in light of this idea.

In fMRI studies of subjects remembering and imagining events, they have identified what they call the ‘core brain system’ which integrates information from past experiences about relationships and associations and uses the information to construct mental simulations. There is a large overlap between the areas of the brain involved in elaborating past and future events. Further there is a large overlap with the default network we use for day-dreaming.

The cognitive machinery outlined by this line of research would convincingly be worth its biological cost by giving us effective and appropriate behaviour. Memory may not be entirely accurate but it is wonderfully useful.
ADDIS, D., WONG, A., & SCHACTER, D. (2007). Remembering the past and imagining the future: Common and distinct neural substrates during event construction and elaboration Neuropsychologia, 45 (7), 1363-1377 DOI: 10.1016/j.neuropsychologia.2006.10.016
Schacter DL, Addis DR, & Buckner RL (2007). Remembering the past to imagine the future: the prospective brain. Nature reviews. Neuroscience, 8 (9), 657-61 PMID: 17700624
Bar, M. (2009). The proactive brain: memory for predictions Philosophical Transactions of the Royal Society B: Biological Sciences, 364 (1521), 1235-1243 DOI: 10.1098/rstb.2008.0310

Lamme model - less emphasis on introspection

According to Victor Lamme, the reason that the study of consciousness is so difficult is that it gives priority to introspection and behaviour so, as a result, we are fooled into thinking that we know what we are conscious of. By adding evidence from neuroscience into the mix, he hopes to understand consciousness. He looks at the components of what we call consciousness and teases them apart: the phenomenal experience, the behavioural control, the access to the experience/ability to report it, the working memory, the attention focus. Then he asks what neural events match which aspects of consciousness.

Vision is the sense that Lamme uses to explain his model. The first cut is made between the iconic memory and working memory. The iconic image has more detail, is short-lived and overwritten by a change of scene, is not useful to cognition; while the working memory image has limited detail, can last for some time and is not overwritten by new events, is useful to cognition. The parts of the iconic image that become part of the working memory are those few which are attended to – so this cut can also be seen as consciousness with attention verses consciousness without attention.

Consciousness is then divided into stages using the neural events associated with consciousness. These events are the fast feedforward sweep (FFS) and the recurrent processing (RP). In the FFS, information flows from the visual cortex forward through the dorsal and ventral paths to the motor and frontal cortex area. This sweep starts with purely visual processing and ends with motor and executive processing. In RP, information flows out horizontally in each area and flows back to lower levels, as far back as the original visual cortex. In other words, there is a bottom-up stream followed by a sidewise and top-down spread.

Lamme’s stages are:

Stage 1 – Superficial processing during the FFS. Processing stops at the visual areas if a stimulus is not attended to and is also quickly overwritten by a masking image.

Stage 2 – Deep processing during the FFS. If stimulus is attended to but is quickly overwritten the result is that processing reaches the prefrontal and motor areas but remains unconscious.

Stage 3 – Superficial processing with RP. If the stimulus is not masked and has time to evoke RP but is not attended to or is neglected (inattentional blindness, change blindness, attentional blink), the FFS does not travel far but is followed by RP in the area it has managed to reach.

Stage 4 – Deep or widespread RP. If the stimulus has time for the FFS to travel fully forward and is attended to, the RP will span all levels from original visual to executive areas. This consciousness can be reported.

With this four stage model, Lamme labels Stages 1 & 2 as not having consciousness, Stage 3 as iconic representation and Stage 4 as working memory representation. The travel of the FFS forward is associated with attention and the spread of RP is associated with phenomenality.

Lamme argues that the inclusion of neuroscientific evidence is important.

“My main objection is against a form of cognitive psychology where mental constructs are taken as undeniable truths to which neuroscience has to be fitted. I would argue that in the study of consciousness, there are no undeniable truths.

That is the standard approach in science. Intuition told us the sun revolves around the earth, while in fact it is the other way around. Intuition dictated creation, where evolution is the counterintuitive scientific answer. To make scientific headway in our science of consciousness, we need to acknowledge that our intuitions may be wrong and need to be set aside. The upshot is that – finally – we may start solving the questions that have been bothering us for the ages.”

What a breathe of fresh air to have introspection taken off its pedestal!

I have my own list of what explanations of consciousness should address and Lamme touches most of them. My disappointment is the sparse mentions of the role of the thalamus and the nature of synchronous activity as well as the lack of a mention of the apparent projection into the near future of the representation of moving objects.
Lamme, V. (2010). How neuroscience will change our view on consciousness Cognitive Neuroscience, 1 (3), 204-220 DOI: 10.1080/17588921003731586


Memories move through stages when they are formed: they are ‘encoded’ or some other process of being prepared (perhaps working memory), they are held in an early stage (short term memory), changes occur in synapses in both the hippocampus and cortex forming a somewhat stable memory from minutes to hours later (synaptic consolidation), the memories are processed so that they are less dependent on hippocampus and more on the cortex from days to years after (system consolidation). Consolidated memories are fairly stable – but- when they are recalled into consciousness, they can be modified. This seems to be the main way in which memories are re-consolidated (or changed) over time. Passing a memory through consciousness means it is re-stored, either unchanged or updated depending on circumstances. Much of sleep is busy with the consolidation and re-consolidation of memories. This is a simplified outline of the current science on memory and it is important to realize that more is unknown than known.

An interesting item in BPS research digest (here) looks at experiments with a treatment for traumatic memories, EMDR, eye movement desensitization and reprocessing.

A controversial treatment for post-traumatic stress disorder involves the traumatised person holding a painful memory in mind while simultaneously following with their eyes the horizontal movements of their therapist’s finger… Raymond Gunter and Glen Bodner have tested three possible explanations…

(first) relative to staring straight ahead, eye-movements increased arousal levels. (undermining) the idea that eye movements activate an innate investigatory reflex that inhibits fear and provokes relaxation.

A second experiment showed that both horizontal and vertical eye movements reduced the vividness and emotionality of the students’ memories. (undermining) the idea that horizontal eye movements aid interhemispheric communication, thus allowing the more rational left hemisphere to process the right hemisphere’s traumatic memories.

(third) experiment showed that the students’ memories became less vivid and emotional, not only when they performed concurrent horizontal eye movements, but also if they instead performed a simultaneous simple hearing task. This undermines the idea that EMDR works specifically by taxing the so-called “visuo-spatial sketch-pad” of working memory. It suggests instead that the mechanism underlying EMDR is a more general effect based on taxing the big boss of short-term memory - the central executive.

…performing a concurrent task, be it eye movements or some other distraction, while also recalling a painful memory, allows a person to be exposed to that memory, without having the mental resources available to get too upset by it. Over time, this process acts like a form of gentle exposure to the memory, as the person learns that they can, after all, cope with their past.

This seems a clear case of a memory being forced to change during re-consolidation – by passing through consciousness under conditions that modify the memory.


I knew a man once that only really thought he understood the meaning of a concept if he knew it through history, fiction or anecdotal narrative. I would not have credited such a way of understanding the world except for knowing him. Some people really ‘get’ algebra and they feel they really know something if they can describe it mathematically. Other people need their explanations to be graphic: an illustration, a map or a diagram. Still others need ideas to be verbal in order to easily grasp them. My friend needed a dramatic plot or a parable to have that feeling of understanding. Words were not enough; there had to be a plot. Most of us can use many or even all those vehicles to understanding, maybe more ways that I have never noticed. We use different ways to understand different things.

Most people I know understand their own lives in narrative form. G. Strawson argues that this is not true of all of us (here). The article starts with this abstract:

I argue against two popular claims. The first is a descriptive, empirical thesis about the nature of ordinary human experience: ‘each of us constructs and lives a “narrative” . . . this narrative is us, our identities’ (Oliver Sacks); ‘self is a perpetually rewritten story . . . in the end, we become the autobiographical narratives by which we “tell about” our lives’ (Jerry Bruner); ‘we are all virtuoso novelists. . . . We try to make all of our material cohere into a single good story. And that story is our autobiography. The chief fictional character . . . of that autobiography is one’s self’ (Dan Dennett). The second is a normative, ethical claim: we ought to live our lives narratively, or as a story; a ‘basic condition of making sense of ourselves is that we grasp our lives in a narrative’ and have an understanding of our lives ‘as an unfolding story’ (Charles Taylor). A person ‘creates his identity [only] by forming an autobiographical narrative – a story of his life’, and must be in possession of a full and ‘explicit narrative [of his life] to develop fully as a person’ (Marya Schechtman).

Strawson identifies four characteristics of this sort of life story narrative. First, the person has a diachronic rather than episodic viewpoint, or in other words, they live in a time continuum rather than in the present. Second, the person has to have a tendency to search for a unifying or form-finding construction. Third, the person uses story-telling conventions. Fourth, the person continues to revise the story.

Like episodic and verbal type memory, narrating a our life seems to me to require that the elements come from conscious experience. If we look at this from the view point of learning in general and learning about ourselves in particular than the reason for narrative seems clear.

  1. We construct and remember experience in the form of memories of moments of consciousness. The memory has the setting, the on-going action and especially important things such as elements that are surprising or significant. To learn from experience, we must have experience.

  2. We remember moments of consciousness in a time ordered sequence, making a little episode of the episodic memory. We can attach meaning to such episodes by associating them with cause and effect links. Causality allows us to use our experiences to understand and predict the world around us.

  3. As memories get older, they are consolidated into larger and larger units. Many nearly identical episodes became one. All the trips to work in a first job, long ago, become one memory. The trip to the store yesterday is still an individual memory. It seems that memory is reworked in light of our present knowledge and interests over and over again.

  4. If we want to tell someone about something that happened, we put these memories in words and those words than become associated with the memory. Often we talk to ourselves about a memory and by doing so make it a narrative to be remembered, partly in a narrative form.

Thus moments of experience become meaningful episodes which become summaries of life and finally become narratives.

Strawson believes that personal narrative is not a universal human characteristic and he may be right. But one thing we know is that long, long ago humans learned to make tools, to harass fire and use language - then, well fed, they sat around the fire and told stories into the night.

Motor actions connected to memory

Neurophilosophy has a posting on embodiment (here) that looks to a motor action/emotional memory link using K. Dijkstra’s work among others.

These results show that bodily movements can influence the rate at which autobiographical memories are recalled as well as the emotional content of the memories. The results of the first experiment demostrate that what we do with our bodies can affect how we think - memory recollection was more efficient when the direction of movement was congruent with the valency of the emotional content of the memory. The second experiment further demonstrates, for the first time, that meaningless bodily movements can also influence what we choose to think about, with upwards movements being associated with positive memories and downward movements with negative ones.

It is well known that memory recollection is facilitated when the context in which recollection occurs matches that in which encoding took place. Classical studies of context-dependent retrieval focused on aspects of the environment in which memory encoding and retrieval take place, and Dijkstra extended this to show that context also includes body posture. The new findings show that movements which are completely unrelated to the encoding of emotional memories can also influence their retrieval. They add to a growing body of evidence that supports the embodied cognition hypothesis; specifically, they provide evidence that thinking involves creating mental simulations of bodily experiences, and that knowledge is represented by partial re-enactments in the brain which activate the same systems associated with real experiences.

Past, present and future

The blog Neurophilosophy has a review of the connection between space and time, and between direction and past or future. (here) It is worth reading in full as I am only talking about a very little part of the review.

We know that imagining a future event is dependent on memory, because patients with amnesia cannot imagine new experiences. It involves piecing together fragments of past experiences to generate a plausible simulation of what might happen.

It seems that memory and imagination/planning use the same basic mental processes. There is the same framework for consciousness of now, of memories and of ‘what if’ scenes. Consciousness is used to create memory, memory is used to produce future imaginings and short term projections are use to produce the consciousness of now. Only what is found in consciousness is stored in episodic memory. We know that stimulations of the future and imaginings are constructed from memory scrapes. And our conscious experience is a slight prediction. The difference between our current consciousness and recalled memories or imaginings seems to be in the amount of detail and the vividness of the experience. I presume that both the detail and the vividness is due to the availability of the actual sensory data in the sensory cortex areas that is only there for a few moments.

Several ways of remembering

ScienceDaily reports on the dissertation of Kristina Kompus at Umea University on the entirely different signal paths for spontaneous (just happened to remember) and deliberate (try to remember) activation of memories. (here)

…these two different ways of remembering things are initiated by entirely different signal paths in the brain. Efforts to retrieve a specific memory are dealt with by the upper part of the frontal lobe. This area of the brain is activated not only in connection with memory-related efforts but also in all types of mental efforts and intentions, according to the dissertation. This part of the brain is not involved in the beginning of the process of unintentionally remembering something as a response to external stimuli. Instead, such memories are activated by specific signals from other parts of the brain, namely those that deal with perceived stimuli like smells, pictures, and words….memories do not need to be emotionally charged to be revived spontaneously, unintentionally. Nor do memories that are revived spontaneously activate the brain more than other memories…The studies also reveal that our long-term memory is more flexible that was previously believed. There is not just one single neurological signaling path for reliving old memories but rather several paths that are anatomically separate. …The dissertation uses a combination of two imaging methods for the brain: functional magnetic resonance imaging (fMRI) and electroencephalography (EEG). The methods yield different information about the function of the brain. By combining them, Kristiina Kompus has been able both to determine what part of the brain is activated and how the activation proceeds over extremely brief time intervals, on the order of milliseconds.