Lecture 26: Consciousness and Intelligence Flashcards
The role of intelligence
- Higher cognitive Processing (reasoning, problem solving, learning) can explain part of our intelligence.
- intelligence is very different from on person to another. It is not just because of IQ, it is different because we have multiple features in our intelligence - we can be srong in one system of intelligence and weak in another.
- Intelligence is about handling things you do not know the answer to, it is about solving problems, taking the information that is from your environment and making something new out of it (action,deecision,recognition).
- The more we are getting into higher complexity and condition (the hardeer the problem/condition), the more noticeable the differencese across individuals are.
Definition of Intelligence
According to Carl Bereiter, intelligence is “what you employ when you don’t know what to do,”.What you do when you don’t know what to do.
* The Gottfredson-organized “mainstream science on intelligence” study places an emphasis on reasoning, planning, problem-solving, thinking abstractly, understanding complicated ideas, picking things up fast, and learning from experience.
* The American Psychological Association (APA) notes that “individuals differ from one another in their abilities to understand complex ideas, to adapt effectively to the environment, to learn from experience, to engage in various forms of
reasoning, and to overcome obstacles by taking thought.”
* Can come with controversity.
What is intelligence?
- Humans are able to detect their surroundings, pay attention to pertinent inputs, memorize episodic and semantic information, and so on. Knowing what is important for us in a specific moment in order to do what you have to do. We have concepts of everything (identify features concept, so that if you see something you have never seen, you can compare it to what you already know). They fit into a concept based on their features.
- adapt our behavior to the environment;
- choose the circumstances that are most suitable
- change the world when adaptation and selection are not a possibility, these activities must be combined in some way. Sometimes we need to restructure what we already learned: create new concept, new connections.
- There is this notion that we have general intelligence. This general intelligence is about coordinating all the subtypes of intelligence. Using what you are goos at and compensating for other things that you are not so performant with.
- We have multiple cognitive abilities and it is the idea thatt they all will communicate togetheer in order for you to achieve your goals.
- it is about handling the different parts.
Dont need to learn this figure
Types of intelligence
- Occipital and temporal areas process sensory information in the first processing stage: the extrastriate cortex (Brodmann areas - BAs - 18 and 19) and the fusiform gyrus (BA 37), involved with recognition, imagery and elaboration of visual inputs, as well as Wernicke’s area (BA 22) for analysis and elaboration of syntax of auditory information. First stage = perceptual information processing.
- Integration and abstraction of the sensory information by parietal BAs 39 (angular gyrus), 40 (supramarginal gyrus), and 7 (superior parietal lobule) correspond to the second processing stage. second processing stage= integrate the info andput it in a basract world.
- The parietal areas interact with the frontal lobes in the third processing stage and this interaction underlies problem solving, evaluation, and hypothesis testing. Frontal BAs 6, 9, 10, 45, 46, and 47 are underscored by the model.
- The anterior cingulate (BA 32) is implicated for response selection and inhibition of alternative responses, once the best solution is determined in the previous stage.
States of consciousness
- States of consciousness depend on how alert we are, how vigilant we are and it depends in what we are doing.
- Different states of consciousness involved different neural networks and different activity.
figure 1: states of consciousness
figure 2: neural activity can vary across stages of consciousness. Use any type of electrophysiology measure to see the differences between each stage.
awake = slow waves, oscillations that are very slow or very fast.
sleeping = activity shifts, we can look at different stages of sleep and identify if someone is in their REM stage based on the neural activity.
Experience of consciousness
- Taking these different aspects of consciousness is what we call the phenomenal realiy. We can have this sesne that is associated with our own thinking of our present, past and future.
- Phenomenal consciousness needs to be accessed. This access of tthe phenomenal consciousness will result in qualia.
Qualia = qualia is different for each of us. I don’t see the colour red like you see the colour red. You will not have the same feeling associated with one colour. What is your subjective notion of what you are conciously experiencing.
When phenomenal consciousness meets access consciousness = qualia. - When you have a stimulus, it can be very weak so maybe you don’t see it or you can have a vibrant, salient stimulus. It is also about attention, if there is a lot of stimuli in our environment and you don’t focus on it you wont see it. Through the matrix of figure 2, you can then study phenomenal consciousness and access consciousness.
Figure 1: we have our conscious self, we are aware that we are and that we are seeing something. We are able to verbalize or think about our present state. Then, there is our past, we are conscious of our memories and the brain has a high ability to predict. We can project ourselves in the future.
Figure 2: different stimulations to test different types of consciousness.
ex: (1) we are not paying attention to a stimulus that is very weak: the stimulus will activate the retina and will activate some visual areas, but it will stay there, we will not be able to have conscious access of this information even though the brain can be activated.
(2) now this attentional resources will be able to go and extract some features within these visual regions and will then activate some higher order regions. Maybe we will have some access to this information.
(3) if we are fully attentive, it will create its feedback connection - bidirectionally across the connectivity between the regions.
(4) strong stimulus but not attentive to it = not activate higher order regions and stay in the sensory regions.
(5) adequate amount of attention for a strong stimulus = activate the whole system and consciously access info.
Perceptual Consciousness
- unconsciousness is more than what consciousnness is.
- Usually, in literature, when you talk about unnconscious processing, you will talk specifically about what we see in animals so detection of danger of motion, survival… it is a primitive instinct.
how can the brain process conscious information?
visual input → retina → thalamus → LGN → V1→ higher order regions
how can the brain process unconscious information?
Retina receives this info →this creates neural acivattion that will be send somewhere in the brain and the idea is that this infromatuon will be sent to other cortical regions: superior coliculus, pulvinar and the thalamus and then go to the dorsal steam. Direct stream that bypasses V1.
Debates:
- where in the brain happens is tthe difference between consciousness and unconsciousness.
- When?
- How?
Perceptual consciousness:
The Global Workspace Theory (GWT)
- This theory says that when you have visual information it will be processed through local regions. This is what we call local processors. They will connect with regions near them.
- The global workspace theory says that in order for us to have conscious access of the information, we need to ignite the system.
- Ignition means that we will activate other networks in the brain. One specific network will be able to ignite other regions.
- This is what we call when there’s this ignition, the global workspace.
What happens when information is proceessed through another pathway that bypasses V1 (if we talk about visual processing).
Perceptual (un)consciousness: Blindsight
Lesion to V1
* if all of V1 is lesioned in one hemisphere, it will create a contralateral cortical blindess.
* RH lesioned, will not be able to see in left hemifield.
* due to a post-chiasmic lesion (lesion after optic chiasm)
* can be caused by stroke or other like surgeries
* Tested a patient with hemianopia so completely blind on one side. If you fixate a room, you will not see anything on one side of the room, do not see half of visual field.
- patient had to fixate the cross and then say what he was presenting to them.
- the person will say i dont see anything, and then they will say give it a try.
- Person will say angry and will have the right answer.
- This means the person is able to process informatuon even though they don’t have access to this information.
- There is a dissociation between the subjective report of the participant (dont see anything) and the objective measure of it.
- will be able to process the information.
Patient with complete removal of V1
- She had a complete ressection of her occipital lobe.
- Completely blind in right hemifield.
- Presented some dots in her visual field to she if she was able to localize these dots even though she wasnt seing. She was kinda good at it.
- There was a correlation between the selected position that she chose and the real position of the flash that was presented on the screen.
- Motion task: presented motion in her blind hemiflied ans she said “i dont see anything” but she was detecting every time the motion. She said “ i am not seing anything but I am sensing something. she could not explain what she was sensing. This is the Riddoch Pennomenon: Phenomenal consciousness that we cannot access to the visual information. So she had sensation of visual emotion in her blind hemifield.
- Activation in brain: when stimulated her normal hemifield, the normal regions were activated. When stimulated her blind hemifield, there were some local processors that were activated. There were some extrastriate regions so higher order visual areas that were activated (pulvinar, superior colleculius) → there was ann extra striate pathway meaningfully. Only specific nodes were activated, did not ignite.
Disorder of consciousness
- Even though we share a lot of everyday activities with other people, we do not have access to their conscious experiences and are only able to grasp them through their self-reports, subjective questionnaire.
- Patients with disorders of consciousness who are conscious but unable to talk or act willfully are unable to tell others about their conscious experiences (coma). In fact, it is sometimes hard to tell whether someone is conscious or not.
- Naci et al., 2014 provide a neural measure that accurately predicted each participant’s conscious experience in a sample of healthy volunteers.
- they looked at all the neural netwroks involved in different systems cognitive abilities and then compared it to patients.
Compared the patients with the controls.
- Presented a movie of called bang (from hitchcock).
They presented this movie to two patients and two controls.
(A) first thing we were looking at is how similar the individuals brains were when looking at the movie. How much of the brain will synchronize together.
(B) resting state - if we don’t do anything are we similar = no
(C) if we look at scrambled images, perceptual images that will not activate a lot of cognition or emotions (only activate visual areas and auditory areas).
Compared A and C: looked a specific components that make us conscious when we see a movie like that.
In this image below (red images): you see different networks. The first network that you see is the brain activation when there is auditory information. able to extract auditory, visual processing and executive processing.
Figure A is the auditory networks that they were able to extract
B is about the visual information (posterior areas that are activated).
Figure C is about the higher order regions executive processing that is activated. These were all in healthy participants
Results (healtjy patients vs coma state)
- healthy group activation: purple = auditory, blue = visual.
- Frontal, parietal networks that are activated in B and C, in healthy groups.
- patient 1 was in a coma state. He did not have any frontoparietal activation so the excutive cognitive abilities of this patient was not activated when he was seeing the same movie as healthy group
- patient 2 was in a coma state aswell but his brain was activated as healthy controls. With the brain acivity, we were able to say that this patient, even though he cannot respond, he is conscious.
Artificial intelligence
- Artificial intelligence was developed through our notions of neuroscience of how the brain works.
- Creating computer models for representing cognitive processes
- Parallel distributed processing (means we can process a lot of info at the same time, multiple inputs that will activate multiple neurons).
- Knowledge can be represented in the distributed activity of many units.
- Weights determine at each connection how strongly an incoming signal will activate the next unit. Am i strongly connected to another unit or weakly connected to another unit = weight of connections.
unit = conceept of neuron
weight = notion of axon
The Connectionist Approach
- Start with random associations.
* More or less true in the brain, although typically
postulated in simulations. - Hebbian Learning rule
* Two cells that are repeatedly activated
simultaneously together have a tendency to become
“associated,” wherein the activity of one cell
facilitates the activity of the other (Hebb, 1949, TL) - Learning
* Each unit receives information via incoming arrows
and sends information via outgoing arrows
* Each connection (or arrow) has a weight assigned to
it, this is the connection weight- It can be positive (excitatory connection) +1
- It can be negative (inhibitory connection) -1
When we are born, we have at least the minimum connections necessary for us to develop other learning.
So this is a little bit like the hebbian learning rule that we already talked about.
- So two cells that are repeated repeatedly, activated simultaneously together, have a tendency to become associated.
- So this is how learning occurs.
- Each unit receives information via income arrows, so via the inputs. And sends information via outgoing arrows.
- So each connection has a weight assigned to it, and the connection weight can be positive or negative, meaning that neuron can excite another neuron or inhibit another neuron.
Units in the connectionist approach
- “Units”
- Input units: activated by stimulation from
environment - Hidden units: receive input from input units
- Output units: receive input from hidden units
It is like when you are a kid and you have never seen a bird anymore. So, you see a bird, you have a random connection of multiple neurons, and then someone will tell what it is and the brain will create a concept associated to the bird. The brain will extract the information of what it is seeing at that point and say it is a bird.
Machine learning
When we alk about aretificial intelligence, there is machine learning in it andin machine learning, there is deep learning. It is just the algorithms that are used in these different subtypes that differs. In machine learning, we have supervised learning, we have unsupervised learning…
machine learning is used to learn, to predict and to improve. Machine has to be able to recognize what it is seing, able to perform as a human.
Machine learning uses is abour feature extraction and itt uses algoritums to group objects, to leanr and to predict and improve.
