Cognitive Neuropsychology
Hoorcolleges aangevuld met boek ‘Introduction to human neuroimaging’
----
HC 1 | Introduction | H1
Leerdoelen cursus
▪ Describe how a range of experimental methods used in cognitive neuropsychology (sMRI, fMRI,
EEG, TMS, PET, single cell recordings, …) work.
▪ Explain the neurophysiological basis of these measurements and the way they inform us about the
functioning of the brain.
▪ Evaluate the relative strengths and weaknesses of these methods.
▪ Evaluate quality and relevance of cognitive neuropsychological research
▪ Design a proper scientific experiment using the discussed techniques.
Voorbeeldvraag
Zonder plaatjes. Je moet zelf weten waar alles ligt
ongeveer.
→ eerste methode die je hierbij zou gebruiken is
PET (cruciaal: dopamine transport). De tweede
methode is TMS (diPFC ligt op het oppervlakte –
cruciaal = causale rol, dat kan je bv. niet bekijken
met fMRI). Met DBS kan je ook causale relaties
bekijken, maar dat gebruik je als een
hersenstructuur dieper in het brein ligt.
Definition cognitive neuropsychology
The study of the relation between structure and function of the brain and specific cognitive functions (e.g.
language, memory, attention, …)
▪ by investigating these cognitive processes in normal healthy people
▪ by investigating the breakdown of these processes in brain-damaged individuals (as a result of
acquired brain damage (stroke, brain tumor) or as a result of a developmental disorder)
o Vroeger alleen focus op brain damage, nu ook bij gezonde mensen onderzoek
Brain enthusiasm
▪ Scientific potential vs. science fiction
o Claim by Verbeke (2009): in 5 years, brain scans will be common when applying for important
jobs
o 2020: brain scans are still not standard practice.
- Why? The imaging methods we have (brain scans) don’t say something reliable / valid
about one single individual.
▪ Brain scans as evidence in court of law, for:
o Personality assessment
o Control of actions
- "Don't blame me, blame my brain."
o Lie detection
- Video: ze gebruiken ‘decoding’. Linkje uit slide nog kijken (brain decoding how scientists
can read your mind)
, - Ze laten zien dat je images from the brain kan decoden (wat mensen eerder gezien
hebben/over gedroomd hebben).
- Dus ze kunnen zien in de hersenen of je al eerder in een ruimte bent geweest bv. Maar
werkt nog niet 100% om echt te gebruiken.
- En als je ziet wat voor image ze daaruit genereren, is het nog een groot verschil van wat
laten zien is.
- Visual cortex is most studied in the brain, so here you maybe can say something more
about what an individual has seen, but to differentiate between a psychopath and
healthy controls might be more difficult due to individual differences.
▪ Brain scans might be overinterpreted by laypersons (liken) > expert witness (getuige-deskundige)
o People will value brain scan evidence much higher than evidence from an expert witness
➔ persuasive power of “neuro-”
Persuasive power of “neuro-“
▪ Ali et al. (2014): How much are individuals ready to believe when encountering improbable
information through the guise of neuroscience?
o Students could easily be convinced that fake neuroimaging instrument (salon hair dryer)
could predict their thoughts
o Students deemed technique highly plausible and were hardly sceptical
Objective diagnosis of diseases?
▪ Good progress for several neurological syndromes
o Brain tumors (saying something about an individual is critical in this case, but also not easy), dementia, mild
cognitive impairment, …
▪ But less progress for psychiatric and mental syndromes
o Depression, autism spectrum disorder, schizophrenia, ...
o Differences on a group level, but not large and consistent enough to allow diagnosis of an
individual subject (huge interindividual differences)
Basis of neural signals
▪ Parts of a neuron: dendritic tree, cell body (or soma) and an axon.
o The cell bodies lay in the grey matter of the cerebral cortex and subcortical structures;
o The white matter contains axons.
▪ Information comes through the synapses to the dendrites and go to axon hillock.
▪ Without input (at rest), cell membrane of a neuron has an electrical potential difference between in-
and outside of -70 mV.
▪ That difference will change depending on the input
the neuron gets. It sums all the input from different
neurons/synapses and across time that information
will effect the potential difference. If that
difference meets a difference of -55mv → action potential.
▪ Post-synaptic potential is determined by integrating input of many synapses at the dendrites. It can
hyper- and depolarize.
,Neural communication
▪ Input neurons (through neurotransmitters): action potentials
over time
→ Membrane potential of post-synaptic neuron depolarizes or
hyperpolarizes
o Input from an excitatory neuron → depolarization =
-70mV becomes -65 mV for example.
- E.g. Glutamate
o Input from an inhibitory neuron → hyperpolarization =
-70 mV becomes -75 mV for example.
- E.g. GABA
▪ Over time, the membrane potential of a post-synaptic neuron
changes in function of the input it receives = signal = a summary
of level of input, the relative degree of excitatory/inhibitory input, when an action potential is
triggered.
Signal description
▪ Simplest signal = sinusoidal oscillation
This signal has a specific frequency.
▪ Frequency: rate of change of signal, e.g. in the time dimension. How much time it takes for the signal
to go up and down. In the time domain it is expressed in Hz.
o 1 Hz = completing a full cycle (going up & down) in one second
o Biological signals never contain just one frequency (happens in artificial signals, e.g. pure tone
- pure tones can only be generated through computer)
▪ Biological signal contain sub-signals or frequency components → Complex signals can be decomposed
into different frequency components. Each component is determined by three parameters:
o Each has a particular frequency (e.g., 1 Hz, 2 Hz, 3 Hz, …)
o Amplitude: how much it goes up and down
o Phase: when it goes up and down
The full signal can be seen as an addition of these three components.
Linksboven: zelfde frequency, maar andere amplitude
Linksonder: amplitude hetzelfde, maar frequency groen
groter
Rechtsboven: phase is anders, frequency en amplitude
zelfde
Rechtsonder: alle drie zijn anders
Dus: je hebt een signal, die kan verschillende dimensies hebben → frequency, amplitude, phase.
En je hebt een ongelimiteerd aantal van frequenties, maar wat je kan meten is niet ongelimiteerd.
Je meet het signaal niet constant, je meet het signaal bv. 1000x per seconde (1000Hz). Maar je hebt informatie
van de gehele cycle nodig om te weten wat voor frequentie je eigenlijk hebt. Als je een hogere frequency hebt,
heb je een beter zicht van hoe veel het op en neer gaat.
▪ Linksboven: als je alleen de bovenste punten van de blauwe lijn meet, lijkt het alsof dat signaal
helemaal niet is veranderd. Dus hoe hoger je sampling frequency, hoe beter je zicht op wat er precies
gebeurt in de tussentijd. Dan heb je een beter zicht op het signal.
▪ Dus the higher your sampling frequency, the better you can describe the signal. But you don’t get all
information, you get half of it because it is limited.
Frequency spectrum: measured range of frequencies (is limited)
▪ Highest frequency
o Limited by sampling frequency (= how often the signal is measured)
, o ½ * sampling frequency (Nyquist sampling theorem) = highest you can measure
o If you have sampling frequency of 100Hz, the highest frequency you can measure is 50Hz.
▪ Lowest frequency
o Limited by how long the signal is measured
o 1 / number of seconds measured
- If the signal goes up and down very slowly, you need a long time to see when the full
cycle has been finished. If you only measured a short time, you don’t have information
about when it goes down and up again.
- To get information at 0.5 Hz you need 2 seconds to get the whole cycle. If you only
measure 1 second, you only get half of the cycle.
- Linksboven: signal goes up and down, so to know something about 1Hz, you should
measure 1second, because that is the time to go up and down.
- 1 / number of seconds measured. Voor 0.25Hz is het bv. 4 seconde.
Filtering: attenuating or excluding certain part of measured frequency spectrum
(low-pass, high-pass or band-pass)
▪ Als je low frequenties niet wil weten, gebruik je bv. high-pass. In
high/pass filtering the lower frequencies are attenuated.
▪ Als je high frequenties niet wil weten, gebruik je low-pass. In low-pass
filtering the higher frequencies are weakened (attenuated) or
completely removed from the signal.
o Also called smoothing.
▪ Als je een bepaalde frequentie wilt bv. gebruik je de band-pass. All
frequencies above and below the range you want to know, are attenuated.
Spectogram: strength of each signal component at each moment in time (right picture)
links: EEG signal with eyes open
= slow wave en small
amplitudes.
Eyes closed: amplitudes get
higher (that is why it becomes
red).
So right is what you see left but then in the time frequency domain.
Molecular and hemodynamic signals
Electrophysiological changes are connected to other kind of changes...
▪ At a smaller scale: movement of chemical substances and molecules
o E.g. depolarization: influx of Na+, repolarization: outward current of K+
o E.g. calcium concentration higher in electrically active neurons → two-photon calcium
imaging (invasive – single neuron resolution) is useful to measure neuronal activity
▪ At a larger scale: hemodynamics (weergave energiebehoefte)
o Blood supply is adjusted to current energy needs
o So if you need more energy, the blood supply will increase
Energy consumption
▪ Electrophysiological events require energy
▪ Amplitude of potential changes are not necessarily best predictor of energy consumption
o Action potential = passive chain of events that does not consume much energy
o Once an acction potential is generated, everything that follows is a passive change of events
(you don’t need much energy for that).
▪ Restoring resting potential requires energy → that is why energy consumption of neuron could
correlate with number of action potentials.
