Functions 1 (1/14)
"functions" in a social milleu 2
- "liberal" and "conservative" parts of brain: not true - fancy statistics ad high-tech data collection can't make up for poorly defined questions - decisions about what to measure and how to measure it are never made in a vacuum - later we will talk about how definitions of cognitive processes were shaped by the goals of agencies that funded foundational work in this area (much funded by military) - basic facts about perception, attention, memory, and decision-making are also shaped by a larger social context, although in less obvious ways than things like political orientation or criminality - actual cognitive functions: recognizing faces, remembering info, using memories, etc. (not political vs. conservative)
double dissociation between Broca's and Wernicke's aphasia
- Broca's aphasia: 1. long pauses between words; effortful 2. words are conceptually related 3. is this an issue with syntax / grammar, putting words into coherent sentence, or something more general? 4. words are memorized, and memory structures are involved with putting the sentences / right words together -> he appears to have difficulty structuring sentence 5. writing ability (depending on patient) can reveal more info 6. do what extent is thought supported by internal monologue? can you not think things if language has no capacity to express them? -> related to philosophy - Wernicke's aphasia: 1. speech does not make sense, appears to not understand what was said 2. repeating stock responses (syntax and words may be separate) 3. fluent speech; clear structure that may or may not be grammatical (breaking language into 2 parts: grammar and lexicon, knowledge of words) 4. patients are aware of deficit, leading to possible loneliness in language-based social interaction
holism
- Fluorens and his contemporaries didn't have great methods to measure cognitive capacities in animals - for sensory and motor abilities, they could demonstrate associations between subcortical lesions and loss of function, so even the "holists" admitted some degree of specialization - but leaning, as Fluorens did, on Descartes, for one's definitions of mental functions and trying to apply this to studies of pigeons and rabbits has some obvious problems - this tension between localism and holism will animate a lot of the debates in modern cognitive neuroscience - didn't have any categories of thought to understand subtle aspects of motor processing; no framework for looking at how different forms of memory might look different / related or attention, and testing for that in animals; cannot localize
phrenology falls out of favor
- Fluorens: early and vociferous critic of phrenology - his experiments (mainly with pigeons and rabbits) established distinct functions of cerebral cortex, cerebellum, and mid-brain structures - but, he failed to find any correspondence between specific regions in cerebral cortex and specific senses - instead, he insisted that the size of the lesion was critical, and that "understanding" or "intelligence" were the result of the cerebral cortex as a whole - this position was widely held by mid 19th century physiologists and anatomists - issue: not well worked-out theory of what mental processes might be like
a holist critique
- Sigmund Freud critiqued the "diagram-drawing" school of aphasia research in his early neurological research - he felt it was naive to expect the brain to be organized according to psychological categories like "word forms" or "concepts" - he proposed instead that language emerged from "fields" connecting perceptual and motor abilities to "drives" -> many synaptic connections leading to fields of complex behavior output (vs. simple flowchart, chain of events) - this is a durable tension in cognitive neuroscience - can we directly relate brain area functions and cognitive functions or not? - is the brain really organized according to our descriptions of cognitive phenomena? - different understanding of what holism can mean -> starts as undifferentiated mass, but fields are organized by connections to and from outside world - back and forth between brain data and does that allow us to revise / refine psychological theories
case studies in cognitive neuropsychology
- concerns with case studies: 1. pre-existing conditions: we rarely have high-quality data about cognitive abilities before a brain injury 2. test validity: cognitive tests can be excellent at the population level, but idiosyncratic data are common (normal range of performance can look different depending on population and can even change over time; time of day, etc.) 3. confirmatory bias: with a single patient, there are limits to how many things one can test; this limits the precision with which the specificity of deficits can be determined (if we wanted to know about how spatial awareness is related to Tan's language processes? -> we can't, so being able to ask questions and form key questions between related processes as an experimenter is important); good scientists use what the data tells them, but we also begin with a set of assumptions that we may not know are true or not 4. the holist critique: for complex cognitive functions, there are many ways they could be impaired, so just demonstrating a relationship between injury to a particular brain region and performance is not enough to ascribe that function to that region
elements of western aphasia battery 2
- consider how this relates to Broca's description of Tan: 1. impaired spontaneous speech and repetition 2. intact comprehension (reading?) 3. modern neuropsychologists think Leborgne was probably apraxic (but would be able to say that with much more confidence today due to diagnostic categories aligning with specific cognitive assays and standardized tests) 4. rather than an impressionistic description, impairments can be described quantitatively 5. this has clear advantages over informal, subjective assessment, but raises its own issues
defining functions part 1
- from early phrenology to early neuropsychology
early attempt at relating structure to function: phrenology 2
- important problems with phrenology: 1. inappropriate anatomical measurement: skull shape not a sensitive measure of brain shape, and anyway the overall shape of the brain is the wrong level of analysis 2. unconstrained definition of functions: for example, studies of prisoners aimed identifying regions associated "criminality", or women with different numbers of children to find the locus of "philopropogentitiveness" - problematic since criminality (for example) is embedded within a social context, and not related to "head" at all - at best, could find parts of brain involved in regulation of executive functions that might make them more likely to engage in unlawful behavior
localization and language (aphasia) 1
- in 1861, neurologist Pierre Paul Broca presented the case of "Tan" who had lost the ability to speak after a brain injury decades earlier - he gave the presentation to the newly established Society of Anthropology, of which he was also the founding president - Tan: 1. could only produce 1 jargon syllable 2. seemed otherwise to be "in possession of his senses" 3. in fact, he seemed to understand what was being said (dissociation between general cognitive abilities + understanding language and being able to speak) - not the first observation of specific loss - understanding of language started more informal and became gradually more sophisticated -specific loss of function -> understanding localization of function within brain (vs. just size of lesion)
localization and language (aphasia) 2
- in the 1870s a number of German neurologists (Lichtheim, Wernicke) began making more systematic observations of aphasia; categorization of research procedure and using neurological patients to understand structure of cognitive functions -- looked at language first - collected records of patients and designed tests to look at their impaired capacities; what parts of language system are affected in order to produce pattern of lost function? - is there a part of brain involved in X but not Y? -> can we find patient with loss of function in X and not Y? - they ascribed specific losses of function to "disconnections" between "centers" - so, a loss of the ability to understand spoken words might be due to a disconnection between "A" and "B" - a loss of the ability to do verbatim repetition may be due to a disconnection between "A" and "M", and so on - this approach of describing "centers" for specific functions is an ancestor of a topic we'll be dealing with in the first few weeks of this course ("modularity") - language provides readily analyzable parts that can be looked at -- early topic used to understand in this framework - Lichteim's house: a (low level auditory processing), A (abstract word form representation), B (concepts), M (motor representation), m (word output processes) - language-related disorders are disassociations between letters - mechanistic view of looking at different parts of brain with different representations, that do different processes, and how they are connected and can be disconnected - conduction aphasia: disassociation between A and M (saying a word that doesn't exist) -> skips the B connection, and makes us wonder what brain architecture looks like for processes -> modularity - Broca: motor output, but no conceptual abilities (damage to M or m) -> then try to isolate injury
passage of time between turn of 20th century and 1950s
- largely ignored in cognitive neuroscience textbooks - possibly related to behaviorism, prior to cognitive revolution - gap in literature of cognitive and areas of related brain functions
summary so far
- modern study how brain gives rise to cognitive functions begins with neurological observations from 19th century, but arguments about localization go back much further - from the beginning, difficulties with this have been clear: 1. patients show us what you can do with damage to a brain area, not what a brain area does 2. lesion site and behavioral profile can be very idiosyncratic - cognitive neuropsychology emerged as a discipline in the mid 20th century, and incorporates strategies to improve its inferential power: 1. double dissociation logic 2. standardized tests, or comparison to a reference group - definition of functions are important in themselves; did not talk about specific brain areas - goals of cognitive psychology / neuroscience: how can we dissociate processes from one another, how can we create tasks?
"double-dissociations" in cognitive neuropsychology 2
- modularity: wanting to identify individual modules with specific processes - procedure is important for determining specificity of process we are interested in - double disassociation logic works whether we know where brain injury is or not -> patterns of behavior as result of brain injury - there are many complex motor activities; within them, speech is the hardest; so is the loss of speech because of a motor issue or if it is hard? -> can find both people who can do motor processes but have impaired speech, and reverse where they can speak but not do other motor processes - speech production and comprehension is served by different systems in brain -> both can be impacted independently (specific part of brain and more broadly as different systems)
"double-dissociations" in cognitive neuropsychology 1
- one strategy for dealing with difficulties of interpretation posed by case studies - logically, if two processes can be impaired independently, they can't be the same thing -> very powerful - problems with this: 1. task decomposition: we want to isolate processes but what we have data on are tasks that entail many processes beyond the ones we're interested in (result is true for this task, data, experiment, stimuli, etc. when the question being asked is are the speech and motor systems independent; think about everything that goes into production and planning of experiment / neuroimaging / behavioral study) 2. "impaired" and "unimpaired" are often relative: patients with brain injuries are rarely "perfect" on any non-trivial task (for example, 90% on a task that normal people get 100% on, mild impairment on task that isn't "impaired") 3. many computational models have demonstrated that double dissociations can be produced in highly interactive systems without separable processes (just because you have demonstrated that they can be independently impacted, there isn't a checked box in the brain to separate the processes completely)
early attempt at relating structure to function: phrenology 1
- premise: 1. different parts of brain are responsible for different processes 2. aptitude is related to differences in brain region size (deep relationship w/ eugenics) 3. brain region size influences skull shape (completely wrong, brain is soft and skull is hard) - therefore: careful measurements of people's skulls will reveal relationships between skull shape and various brain functions - no one believes in phrenology anymore - which of these premises are actually completely and obviously wrong? - physiogamy: cousin to phrenology, looking at faces and see what we can learn from them - phrenology as a cautionary tale since it is misguided - any skull morphology differences depending on ethnicity nave no functional consequences
"functions" in a social milleu 1
- some of the problems of phrenology (and physiognomy, the study of facial features) are endemic to doing any kind of scientific research on humans - concepts like "criminality" and "philoproprogenitiveness" are complex constructs that are unlikely to have a simple biological basis - if this seems like a weird anachronism, consider that there are AI systems out there right now that are trying to classify people in various ways based on pictures of their faces - even though we have better tools and techniques for analysis today, we still need to think about how to construct questions about brain activity - facial recognition software being used in criminal proceedings; look at facial characteristics of criminals vs. non-criminal w/ digital imaging facial recognition software - can critique these data sets in technical terms and show how they are problematic (article) - people may still try to use unfounded, poorly justified concepts like this to study facial structure / brain
elements of western aphasia battery 1
- spontaneous speech: fluency and info content of a subject's speech as they answer questions and describe pictures - auditory comprehension: yes/no questions, identification of objects and pictures, and execution of commands and controlled complexity - repetition: repetition of words, phrases, and sentences - naming: object naming, word fluency, sentence completion, and responsive speech - reading: comprehension, oral reading, and stimulus association - writing: dictation, copying, writing of overlearned series, and spontaneous written formulation - praxis: assessing both limb and oral apraxia using simple and complex gestures - construction: assessing drawing, block design, and numerical calculation (includes Raven's Coloured Progressive Matrices; Raven, 1956) -- non verbal IQ measure baseline, spatial intelligence was original design, but used here to get reading on general cognitive impairment levels (along w/ language impairment) - many of these abilities fit neatly in Lichtheim's "house" - the difference is that "deficits" are defined quantitatively, relative to a comparison sample - Broca and Wernicke could look at fluency test scores based on these criteria and adjust based on demographics / educational background
tools for modern neuropsychology
- standardized tests: permit quantifiable statements about degree of impairment relative to a comparison sample - structural neuroimaging: permits in vivo identification of lesion (used to only be able to look at tissue post-death); allows report of person's behavior + also where their injury is - tools are more rooted in psychology than neuroscience - neuropsychology: study of people with injury to brain that impairs some aspect of function (focus on what function is disrupted)
define "normal"
- the "normal" range of performance is a bit of an artifact - sorting and ranking populations this way in general has creepy history - it is increasingly common for researchers to compare patients to a sample of individuals matched to them on a range of relevant factors (region, ethnicity, education, work history) - difference from a standard isn't necessarily a deficit that needs to be treated
