The core problem in post-traumatic stress disorder (PTSD) is memory. Memory causes permanent changes in behavior as a result of prior experience. Without memory, we would be pure automatons, robots unable to benefit from experience, doomed to repeatedly burn our hands on hot stoves, step on the garden rake and get smacked in the face by the handle, and chase balls out into busy streets. Fortunately, we have specialized cells, called neurons, that are extremely plastic, and can literally shape themselves to the environment. Without getting into the very complex issue of brain development, let’s simply say that you learn throughout life, even in the womb. For example, a post-natal rat that developed in amniotic fluid laced with lemon extract will later learn to perform tasks to obtain that scent again. A child that thinks of chasing a ball into the street will pause and be side-tracked at the memory of the stinging rebuke given by a concerned parent. Life has no meaning without memory.
Let me rephrase that. Life could have meaning without memory, but not in the way that we think about meaning. For example, if you were that robotic child chasing balls into the street, and were struck by a car, it could very well be that you would feel exquisite pain. That would be meaningful. You could be designed to feel pain, enjoy sex, and have an extraordinary palate. As occurs in the severe dementia of Alzheimer’s disease, you would simply never remember any of it. Not even a minute later. So while value, e.g., pain and pleasure, might be inherently built into your system, and might have some adaptive value, it would be meaningless in the sense of what we consider meaning, layers upon layers of rich webs of associative chains interwoven into complex cascades of egocentric (self-centered) and allocentric (non-self, or other-centered) meaning. This autobiographical meaning is so complex that one can relate or generate millions of memorial narratives about one’s self, experiences, feelings, relations to others. One can listen to a friend’s phone call (otherwise arbitrary auditory signals) and discover that their child ran into the street to chase a ball, was struck by a car, and didn’t make it; then feel ineffable sadness about the loss of that child’s potential autobiographical record. Without the ability to weave an associative web of meaning, that phone call would otherwise seem like arbitrary, if patterned sound. Sound and fury, signifying nothing.
Oliver Sachs related a story of a patient who developed alcohol-induced Korsakoff’s dementia as a young man in the Navy. He stopped learning at age 19. Even at the age 56 he still thought he was 19 and in the Navy. When Sachs showed him a mirror, the man screamed in horror at his aged image. Wondering about the ethical implications of horrifying this man, Sachs later thought, Well, yes he was horrified, but then he immediately forgot about it. There was no trauma sustained. This is not the case with normal humans.
Pavlovian Conditioning
In 1901 at a meeting of the American Psychological Association, a man named Twitmeyer demonstrated to the top thinkers of the day, including William James, brother of Henry James, how the knee-jerk reflex could be elicited simply by faking a hammer blow to the tendon beneath the knee-cap. The mallet never actually touched the knee, never distorted the tendon beneath, never distorted the stretch receptors in the tendon, and never directly excited the monosynaptic reflex that causes the leg to jerk. Despite being immersed a rich history of associative thought, everyone stared at him blankly, apparently unimpressed. A few years later, Pavlov began publishing on conditioned digestive reflexes. So, perhaps fortuitously, instead of Twitmeyerian conditioning, we now have Pavlovian conditioning. (Due to his unequaled stature, Pavlov was an outspoken, and untouchable political maverick. Nevertheless, it is my understanding that he once said, "There is no revolution ‘out there.’ The revolution is ‘in here.’"
Now, it seems that about everyone in the world knows about Pavlov’s famous experiments, to the point where one virtually (and reflexively) trots out jokes about bells and whistles and drooling, possibly even without understanding what exactly they are talking about. The beauty of Pavlov’s (and Twitmeyer’s) demonstration was the spareness of the procedure. When studying something as complex as a brain (bazillions of neural elements, each having bazillions of connections, how does one even begin to identify what are the crucial elements of memory?) Pavlov, who was studying digestive reflexes, noticed that his dogs would begin drooling when someone entered the room to feed them, which he then termed "psychic secretions." He found psychic secretions fascinating, and then developed procedures to study them, procedures so spare that individual elements of the procedure could be readily identified, reproduced, manipulated, and which suggested the fundamental nature of associative learning. So, while people may think that getting a dog to droll upon hearing a bell, a tone, or a metronome as some sort of trivia, its meaning runs deep. Wallace Stevens once wrote that "Death is the Mother of all Beauty," which sounds about right to me, provided you can associate death with other meaningful aspects of life. And this was Pavlov’s goal.
What is Noticed Becomes a Signal for What is Being Done.
Pavlov identified a handful of key elements in associative conditioning. First, the organism has to have an innate ability to be able to make a response, such as a knee jerk, a salivatory excretion, a startle, and so on, which he termed the "unconditioned response." (the UR). Second, some biologically significant stimulus, such as a hammer blow to the knee, food touching the tongue, or a sudden sharp noise, has to provoke that response. He termed this stimulus the "unconditioned stimulus." Thus, the organism has been selected and designed by nature to have innate responses that correspond appropriately to events in the world. Touch a baby’s cheek, and his head turns in that direction, as if searching for a nipple. No learning required. So far this is like our automaton robot that can pull his hand from a hot fire, but won’t remember not to put his hand in the fire again, much like George Bush going to war.
Unconditioned Stimulus (US) --> unconditioned response (UR)
Hammer to tendon --> knee jerk (postural adjustment to unexpected change in load)
Food on tongue --> salivate (to digest and swallow)
Sharp noise --> startle (jump or freeze, increase heart rate, pump adrenaline...)
While it is adaptive to have meaningful responses to meaningful stimuli, it is even more adaptive to be able to predict and anticipate when something meaningful is going to happen. If there are signs that an event is imminent then one can potentially avoid or facilitate that impending event. If the sight or smell or sound of the fire reminds you of pain, pull you hand from the fire before it burns. And here is where arbitrary stimuli become attached to one’s value register, one’s innate or unconditioned responds.
Pavlov realized that just about any stimulus could be used as a signal for impending biologically significant events. So, before applying meat powder directly to the dog’s tongue, he would signal the event by letting, say, a metronome tick. Normally, a ticking metronome might cause you to turn your head to notice it (an orienting reflex), but it will not cause you or a dog to salivate. However, after following the ticking metronome with the application of meat powder to the tongue, soon the metronome alone will increase salivation in preparation for the meat powder. Now, the once arbitrary stimulus that once only provoked a "noticing response" has been transformed into a predictive stimulus that carries meaning, what Pavlov termed a "conditioned stimulus," capable of evoking a "conditioned response." Now imagine the following sequence of events happening in time, with time flowing from left to right:
On the very first trial of CS--US pairing:
Metronome --> orienting response (noticing)
..................................Meat powder --> salivation
After CS--US pairing:
Metronome --> salivation
Now, armed with this ability to attach meaning to arbitrary sensory stimuli, our robot smells the smoke, hears the snap and roar, feels the heat, and keeps his distance from the fire. Meaning is acquired through association. What is noticed becomes a signal for what is being done. Because neurons that fire together wire together.
Pavlov also noticed that in addition to being able to predict when events are going to happen, organisms can also learn to predict when events are not going to happen. For example, if he continued presenting the ticking metronome but no longer followed it by food, the conditioned salivatory response would begin to fade, a process he referred to as "extinction." What he meant is not that the memory was gone or forgotten, as others have since wrongly postulated, just that it was subsequently modified, overlain with an additional inhibitory association. He went on to develop procedures to demonstrate "conditioned inhibition." For example, if stimulus A predicts shock, the animal will freeze in the presence of A. However, if there are trials in which stimulus A is presented with stimulus X , and no shock occurs, then the animal will inhibit freezing in the presence of the AX compound.
I will have more to say about extinction and inhibitory conditioning in later diaries, because these related process are believed by many to be keys to solving the some of the problems associated with PTSD, which many view as a failure to extinguish conditioned responses, or a failure to overlay former memories with sufficient, new inhibitory learning.
In the meantime, there is a great deal known about how specific sensory pathways meet with particular response pathways, especially for critically important types of learning, such as fear conditioning. Let’s take a quick peek under the hood at some of these circuits known to be critically involved in fear conditioning.
The Brain & Fear Circuits
Many researchers have employed Pavlov’s methods and ideas to discern what neural systems are involved in fear conditioning, which is believed to form part of the basis for anxiety disorders, phobias, depression, and PTSD. Insofar as these circuits have been delineated, there is an amazing correspondence between pathways controlling stimuli, responses, and associations between stimuli and responses and Pavlov’s conceptual framing of unconditioned and conditioned stimuli and responses.
To keep things simple, let’s go along with a great neuroanatomist’s (Larry Swanson’s) contention that "the brain has 3 parts." In the image below, the brain has a green part (cortex), a red part (cerebral nuclei), and a blue part (brainstem). For the most part, all sensory information (except odor, which goes directly to cortex) flows up through the brainstem (thalamus) which sends info to relevant sensory areas of cortex (visual, auditory, etc.). Then through complex descending projections (lower panel) the cortex can stimulate highly differentiated patterns of outputs. The cortical ouputs are primarily excitatory, whereas the cerebral nuclei (sub-divided into striatum and pallidum) are inhibitory. The way they stack up, it functionally and very flexibly gives various patterns of excitation, inhibition, and disinhibition through all the more minor geographies of the brain we are not discussing.
The cortex is like a giant coincidence detector, and it spends a good deal of time synchronizing, correlating, and comparing. So, a lot of associative activity takes place in cortex. This is not to say other areas are not plastic. Au contraire. However, activity in lower regions, such as striatum and pallidum, are associated with motivation and emotion, whereas the brainstem has both primary sensory and motor components. The sensory component (thalamus) directs incoming calls to other appropriate areas of brain. The motor part of the brainstem is chock full of pre-organized motor pattern generators, such as the coordination of neuroendocrine stress responses and autonomic (heart rate, blood pressure, body temperature) outflows. These "low level" sensory and motor responses are so highly organized that animals can survive just fine (under controlled conditions) when the brainstem is disconnected from higher structures. All basic regulatory responses for survival (eating, drinking, and sleeping) are left intact. They are basically automatons that wouldn’t last long in nature, just like the "nineteen-year-old" with Korsakoff's dementia under a doctor's care.
The Amygdala
As an aside, it’s interesting that odor goes straight to cortex, bypassing the sensory thalamus, because the place in cortex it goes to is the cortical amygdala, one of the prime locations of fear conditioning. Odor is very primal. Anyway, the amygdala has long been known to have a role in emotional responses and fear and more recently it has been implicated in positively valenced memories, as well. Animals and people with damage to the amygdala have a lot of difficulty learning to be normally fearful of new things and other people. The main output nucleus of the amygdala is the central nucleus, a striatal component that projects to downstream pallidal and brainstem motor output structures, that give rise to freezing, running, jumping, and squealing, and increases in body temperature, heart-rate, and bloodpressure, all of which can be conditioned through Pavlov’s method of presenting two stimuli together.
For example, a rat might be presented with an arbitrary auditory stimulus such as a tone of a certain frequency followed by an unconditioned stimulus, such as shock. Even after a single such pairing of tone—shock, the rat will subsequently freeze upon hearing the tone.
CS-US Convergence in the Cortical Amygdala
In order for associative learning to take place, the neurons signaling the arbitary conditioned stimulus, in this case a tone, must meet up, or converge with neurons carrying information about the unconditioned stimulus, in this case the pain of electrical shock. Pain is transmitted from the spinal tract to the medullary and pontine brainstem and thalamus and onto the parts of the cortex responsive to pain. Notice that there is a direct projection from the pontine brainstem to the striatal amygdala, for fast, reactive and unconditioned responses to pain.
Now, let's look at the pathway for the tone. The tone travels from the ear up to the auditory thalamus where it relays to the auditory cortex, both of which then project to the lateral (cortical) amygdala.
Thus, both the CS and the US converge on the cortical aspects of the amygdala. When the adjacent neurons fire in succession they strengthen one another synaptically, so that the next time the tone alone fires, the output from the cortical amygdala on to the central nucleus is enhanced. Different parts of this system have been "neutralized" with various drugs at different points in time to demonstrate that the cortical amygdala is indeed a seat of memory acquisition.
I'll simply show one graph (LeDoux (2000), based on G. Quirk) showing the induction of amygdalar activity over the course of repeated trials of fear conditioning. This graph below shows the increase in activity of cortical amygdala neurons during conditioning (on probe trials with tone presented alone without shock) and during extinction, when the tone is presented always by itself without shock.
Notice two things. First, due to the synchronized activation of tone and shock inputs during conditioning, repeated pairings result in ever increasing activation of amygdalar cortical neurons, which fades away when the tone is presented repeatedly all by itself. The amygdala seems infinitely plastic to repeated exposures to fear, but then shows extinction upon removal of the unconditioned stimulus. Second, notice how in between the final block of acquisition trials and extinction, the activity has yet increased even more in the first round of extinction. An increase in plasticity in the absence of further tone--shock pairings. Interesting. More later on this latter observation.
Has this any relevance for human learning?
From Joseph LeDoux's (2000) paper in Annual Reviews of Neuroscience:
THE HUMAN AMYGDALA
Over the past several years, there has been an explosion of interest in the role of the human amygdala in fear. Deficits in the perception of the emotional meaning of faces, especially fearful faces, have been found in patients with amygdala damage (Adolphs et al 1995, Calder et al 1996). Similar results were reported for detection of the emotional tone of voices (Scott et al 1997). Furthermore, damage to the amygdala (Bechara et al 1995) or areas of temporal lobe including the amygdala (LaBar et al 1995) produced deficits in fear conditioning in humans. Functional imaging studies have shown that the amygdala is activated more strongly in the presence of fearful and angry faces than of happy ones (Breiter et al 1996) and that subliminal presentations of such stimuli lead to stronger activations than do freely seen ones (Whalen et al 1998). Fear conditioning also leads to increases in amygdala activity, as measured by functional magnetic resonance imaging (LaBar et al 1998, Buchel et al 1998), and these effects also occur to subliminal stimuli (Morris et al 1998). Additionally, when the activity of the amygdala during fear conditioning is cross correlated with the activity in other regions of the brain, the strongest relations are seen with subcortical (thalamic
and collicular) rather than cortical areas, further emphasizing the importance of the direct thalamao-amygdala pathway in the human brain (Morris et al 1999). Other aspects of emotion and the human brain area are reviewed by Davidson & Irwin (1999), Phelps & Anderson (1997), Cahill & McGaugh (1998).