Az egyéni látásmód idegrendszeri háttere Vidnyánszky Zoltán Neuro-Infobionikai Kutatócsoport, MTA PPKE SE, Budapest MR Kutató Központ, Szentágothai J. Tudásközpont és Semmelweis Egyetem, Budapest 2011
Tartalomjegyzék A látórendzser és a vizuális információfeldolgozás alapjai Aktív látás Plaszticitás, tanulás
Visual Field Representation The axons of ganglion cells exit the eyes via the optic nerve, partially cross at the optic chiasm, and form two optic tracts, so that the right and left hemifields reach the left and right hemispheres. Each optic tract looks at the opposite hemifield, combining inputs from the ipsilateral temporal hemiretina and the contralateral nasal hemiretina. Optic Nerve: 1.2 million ganglion cell axons (myelinated optic nerve fibers)/nerve. Optic Chiasm: partial decussation of optic nerve fibers (nasal fibers project contralaterally, temporal fibers project ipsilaterally). Thus, left side of brain looks at right visual world; right side of brain looks at left visual world.
Visual Pathways The retina projects to four subcortical regions in the brain: 1)Thalamus 2) Superior Colliculus 3) Hypothalamus 4) Pretectum
Retinotopy of V1 Source: Tootell et al., 1982 Retinotopy: Adjacent parts of the visual scene are mapped to adjacent parts of the cortex
Ocular dominance columns seen w/ radiolabled proline Cells in layer IV and simple cells in layer III receive input from only one eye and therefore, respond only to stimuli from one retina (i.e. monocularly driven). Cells in other layers generally receive input from both eyes (i.e. binocularly driven). Size of column = 0.5mm.
Receptive Fields Place an electrode near a cell in a monkey brain Make the monkey stare at a fixation spot without moving its eyes Stimulate various regions of visual space A cell will respond to stimulation in one part of space more than any others The region of visual space that drives a particular cell forms its receptive field (RF) Source: Gazzaniga, Ivry & Mangun, 2002
Receptive Fields Primary visual cortex Simple Cells: Layers III Monocularly driven. Complex Cells: Layers II, III, V and VI. Binocularly driven.
Orientation selectivity
LGN cell-to-simple cell
Simple cells-to-complex cell
Receptive field of complex cells
Magnocellular, Parvocellular and Koniocellular input to V1
Magno vs. Parvo Magnocellular Large cells Colour insensitive Low resolution Fast, transient More sensitive at low contrast Quick and dirty Parvocellular Small cells Colour sensitive High resolution Slow, sustained More sensitive at high contrast Slow and detailed
Lateral Geniculate Nucleus Lesions restricted to the parvocellular layers severely disrupt the processing of color, while lesions of the magnocellular layers leave color vision unaffected. Lesions restricted to the parvocellular layers severely disrupts the processing of fine detail, while lesions of the magnocellular layers leave fine detail vision unaffected.
Orientation columns Preferred orientation of all cells in a column are similar. The preferred orientation of cells gradually rotates as one moves from column to column across the cortex. 1 mm represents 180 degrees.
Hypercolumns 2mm x 2mm x 2mm
Beyond V1 primary visual cortex (V1) extrastriate cortex Macaque Brain Flattened Cortex Source: Van Essen et al., 2001
Visual cortex Macaque monkey Over 30 visual areas Complex organization Visual areas make up ~40% of monkey brain Source: Felleman & Van Essen, 1991
What is a Visual Area? Function Architecture Connectivity Topography Evolutionary Relationships
Functional Definitions an area has a specific pattern of responses (e.g., electrophysiology, fmri response) to different stimulation (visual properties here) disruption of that area (by lesioning, cooling, transcranial magnetic stimulation) affects processing of a particular type of stimuli stimulation of the area (e.g., electrode stimulation) evokes a particular percept or response
Descartes, De l Homme (1632) Tobozmirigy- ahol a test a lélekkel találkozik
Gall frenológiája (1800)
Broca-área (1861) Motoros beszédközpont
Cytoarchitectonics: Definition the use of microscopic methods to examine differences in the cellular organization between cortical areas Korbinian Brodmann, 1909 identified 52 cortical regions popular but outdated scheme
Connectivity Different areas have different patterns of connections with other areas Inject tracers into an area and see which other areas connect to it The number of intermediate connections between an area and V1 can indicate where it lies in the hierarchy of areas
Tools for mapping human areas Neuropsychological Lesions Temporary Disruption transcranial magnetic stimulation (TMS) Electrical and magnetic signals electroencephalograph y (EEG) magnetoencephalograp hy (MEG) Brain Imaging positron emission tomography (PET) functional magnetic resonance imaging (fmri)
horizontal meridian (HM) Retinotopic visual areas have a map VM HM VM calcarine sulcus VM HM VM vertical meridian (VM) left occipital lobe right occipital lobe Each visual area contains a map of visual space polar coordinates eccentricity runs posterior-anterior phase runs superior-inferior The map for visual area V1 lies along the calcarine sulcus Source: Jody Culham
horizontal meridian (HM) Map boundaries divide visual areas Consider just one hemifield/hemisphere vertical meridian (VM) VM HM VM HM VM HM VM left occipital lobe calcarine sulcus Adjacent maps are divided by specific boundaries (in vision, the horizontal and vertical meridia) Source: Jody Culham
Retinotópia Forgó tortaszelet a háttérben gyorsan villogó sakktáblamintával Táguló gyűrű a háttérben gyorsan villogó sakktáblamintával 28.8 másodperces ciklusidő, 12 ciklus egy menetben
Retinotópiás térkép A bal látótér a jobb félteke occipitális lebenyére képeződik le Baloldali látótér Jobb félteke felfújt képe Kilapított nézet
Retinotópia látóközpontok Tootell & al, TICS 1998
Stages / levels of visual information processing Low level Local feature detection (V1 - V3) Intermediate level grouping and segmentation (V4, MT/MST) High Level Object recognition / biological motion (IT, STS, PC)
Selectivity for visual features in early and intermediate visual cortical areas
Functions of the early and intermediate visual cortical areas
Feature specific visual cortical areas Motion-specific area MT+/V5 Moving vs static stimuli univariate analysis Grill-Spector & Malach, 2004
Human visual cortex Category-specific areas (e.g. faces, body parts) Grill-Spector & Malach, 2004
Schematic atlas of human visual cortex Hierarchy vs specialization Grill-Spector & Malach, 2004
Two visual pathways The two visual processing streams for different visual percepts: What (ventral stream)- object recognition main input from slow and detailed parvo system Where (dorsal stream) - spatial perception main input from quick and dirty magno system Source: Mishkin & Ungerleider, 1982
What vs. Where Mishkin & Ungerleider, 1982 what Lesions in IT cortex produce deficits in shape discrimination tasks Gross et al, 1973, Mishkin 1982 where Lesions in parietal cortex produce deficits in landmark task (Pohl et al. 1973)
What vs. How Goodale and Milner, 1991 dichotomy should be what (ventral stream) vs. how (dorsal stream) dorsal system has strong input to motor systems and is essential for using visual information to guide actions
Patient DF: no form visual perception Patient DF has a ventral stream lesion Object agnosia Cannot identify line drawings of common objects Cannot copy line drawings Can draw from memory as long as she doesn t lift hand from paper
Patient DF: acting without perceiving DF Control Posting task Perceptual matching task: performs poorly Posting task: performs well, begins to rotate card in the correct direction when movement begins
Patient DF: acting without perceiving Grasping task DF picks up the objects at stable grasping points (as do normals) Perceptual task even though she is at chance when shown two of the objects and asked whether they are the same or different
DF: limitations to acting without perceiving T task: DF was able to grasp the object at the correct point but half the times made a 90 o mistake in orienting the object toward the slot.
Patient RV: visuomotor deficits but spared perceptual abilities Dorsal lesion Ventral lesion Patient RV has dorsal stream damage (posterior parietal cortex) Object ataxia: unable to use visual information to reach out and grasp objects no difficulty in recognizing or describing objects
Perception vs. Action Vision for Perception Ventral Stream Vision for Action Dorsal Stream Inferotemporal Cortex Posterior Parietal Cortex Object-based Viewer-based Object identification Movements Conscious Automatic Goodale, M.A. & Milner, A.D. (1992). Trends in Neurosciences 15: 20-25.
Aktív Látás
A Látás Nagy Illúziója - a valósághű agyi leképezés révén a látótér egészét egyidejűleg, részletgazdagon látjuk
A vizuális inputnak mindig többféle értelmezése lehetséges, nem lehet visszavetíteni egyértelműen a fizikai forrására, nem működik az inverz optika. A látás értelmezés, nemtudatos következtetések levonásának folyamata azzal kapcsolatban, hogy mi is található környezetünkben (Helmholtz, 1866)
Így van ez a tárgy fényességének meghatározásánál is
Nincs rá kapacitás Cselekvésünk hatékonyságának elengedhetetlen feltétele, hogy csak a szándékaink szempontjából fontos szenzoros ingereket dolgozzuk fel, hiszen a rendelkezésre álló szenzoros és motoros erőforrásaink korlátozottak Az információ nagy része redundáns vagy aktuális szándékaink szempontjából irreleváns
Az aktív és adaptív látás idegrendszeri folyamatai
A vizuális inputnak mindig többféle értelmezése lehetséges, nem lehet visszavetíteni egyértelműen a fizikai forrására, nem működik az inverz optika. A látás értelmezés, nemtudatos következtetések levonásának folyamata azzal kapcsolatban, hogy mi is található környezetünkben (Helmholtz, 1866)
Így van ez a tárgy fényességének meghatározásánál is
Limitált feldolgozási kapacitás Cselekvésünk hatékonyságának elengedhetetlen feltétele, hogy csak a szándékaink szempontjából fontos szenzoros ingereket dolgozzuk fel, hiszen a rendelkezésre álló szenzoros és motoros erőforrásaink korlátozottak Az információ nagy része redundáns vagy aktuális szándékaink szempontjából irreleváns
Az evolúció során kialakuló biológiai látás egy aktív, a rendelkezésre álló információt szelektíven és az egyedfejlődés során szerzett tapasztalat alapján feldolgozó folyamat funkciója: Az aktuális szándékok szempontjából, valamint a tapasztalat alapján legfontosabb vizuális információ kinyerése A cselekvés vizuális irányítása
Aktív látás [Aloimonos és mtsai, 1988; Ballard, 1991; Findaly és Gilchrist, 2003] alatt elsősorban a látás kiválasztó funkcióit értjük, melyek az aktuális szándékok és a tapasztalat szerint befolyásolják, hogy egy adott pillanatban a környezetünknek melyik része kerül a látómezőnkbe és hogy azon belül is mi milyen hatékonysággal fog feldolgozódni. Mechanizmusai: szemmozgás figyelem tanulás
Funkció A vizuális figyelmi szelekció A környezetétől leginkább eltérő vagy az aktuális célok szempontjából legfontosabb vizuális tárgyakkal, eseményekkel kapcsolatos információ kiválasztása. A kiválasztott információ feldolgozásának hatékonysága megnő míg a figyelmen kívűlié gátlódik. Típusai stimulus-vezérelte (bottom-up) szándék-vezérelte (top-down) A figyelmi szelekció egységei a látómező specifikus területei vizuáis tulajdonságok (szín, mozgás stb.) vizuális tárgyak
Attentional networks Alerting Sustained attention, to increase and maintain response readiness in preparation for an impending stimulus. Structures: frontal (DLPFC, ACC) and parietal (IPC) regions, right hemisphere tonic and left hemisphere phasic alerting; Orienting/Selection The ability to select specific information from among multiple sensory stimuli; exogenous and endogenous orienting Structures: pulvinar, superior colliculus, superior parietal lobe, temporoparietal junction, superior temporal lobe and frontal eye fields Executive functions Includes supervisory functions, conflict resolution and focussed attentionstructures: frontal cortex (DLPFC, ACC)
Visual Attention We can think of eye fixation as mechanical pointing device, and localization by attention as a neural pointing device. Thus one can think of vision as having either mechanical or neural deictic (pointing, showing) devices: fixation and attention. Attention is a pointer to parts of the sensorium that is manipulated by current task goals (Ballard et al, 1997) close connection between eye movement and attention overlapping neural network: FEF, precentral sulcus intraparietal sulcus
Saliency map - bottom-up
Top-down attentional selection Biased competition model [Desimone & Duncan, 1995] Top-down multiple stimuli in the visual field automatically engage in competitive interactions figyelem attention can biase the competition in favor of the attended stimulus FS versengés FKS as a result, processing of the attended stimulus is enhanced while ignored stimuli are suppressed. Bottom-up gátlás FS - megfigyelt stimulus serkentés FKS figyelmen kívüli stimulus
GOAL-DIRECTED AND STIMULUS-DRIVEN ATTENTIONAL SYSTEM Dorsal goal-directed attentional network is involved in preparing and applying goal-directed (top-down) selection for stimuli and responses (intraparietal cortex and superior frontal cortex) Ventral stimulus-driven attentional network is not involved in top-down selection. Instead, this system is specialized for the detection of behaviourally relevant stimuli, particularly when they are salient or unexpected. Attention systems: Dorsal - blue, top-down, (rightward bias) Ventral - orange, stimulus-driven, (reorienting deficit) Corbetta & Shulman, 2002
Neural effects of attentional selection Biased competition Raynolds et al,, 2004
Neural effects of attentional selection O Craven et al,, 2000
Long-lasting effects of attentional selection Perceptual learning in adult humans and animals refers to improvements in sensory abilities after training.
Model of perceptual learning (Seitz & Watanbe, TICS, 2005) Long-term sensitivity enhancements to task-relevant or irrelevant stimuli occur as a result of timely interactions between diffused signals triggered by task performance and signals produced by stimulus presentation.
however in most complex natural scenes, an ideal observer should also attenuate task-irrelevant sensory information that interferes with the processing of task-relevant information (Ghose 2004, Vidnyánszky & Sohn, 2005) Goal: to test the hypothesis that learning leads to suppressed perceptual and neural responses for task-irrelevant information, which competes with the processing of the task-relevant information during training.
Gál et al., 2009 Filtering out visual distractors
perceptual motion sensitivity (coherence threshold) Learning effects on the perceptual motion sensitivity Task: Motion coherence detection Which interval contained coherent motion? 1 0.8 attended during training neglected during training control 0.6 Before training, there was no difference in motion detection thresholds for the two directions that were present during training 0.4 0.2 After training the motion coherence threshold for the task-relevant direction was significantly lower than the threshold for the task-irrelevant direction 0 before training sessions after Gál et al., (2009)
Learning effects - fmri Before training: no difference between the fmri responses evoked by the two motion directions After training: task-irrelevant direction evoked significantly smaller fmri responses than task-relevant direction Gál et al., (2009)
Learning effects - ERP Grand average ERP responses shown for the PO8 (A-D) and Pz (E- H) electrodes. There was no difference between the ERP responses to the task-relevant (A,E) and task-irrelevant (B,F) directions before training. After training, the magnitude of motion signal strength dependent modulation of the ERP responses in the 300-550 ms time interval is reduced in the case of taskirrelevant direction (D,H) compared to that in the case of task relevant direction (C,G). Gál et al., (2010)
Learning effects on the motion strength dependent modulation of the ERP responses Two different components of learning effects on motion coherence dependent ERPs: ~ 300 ms after stimulus onset computed within a cluster of occipitotemporal electrodes, which was primarily due to the training-induced modulation of ERP responses to the direction that was task-relevant during practice. ~ 475 ms after stimulus onset. computed within a cluster of parietal (B) electrodes, where learning affects primarily the processing of the direction that was task-irrelevant during training Gál et al., (2010)
Az hosszú távú tanulás idegrendszeri hátterét képező folyamatok A tanult szenzoros ingerek vagy motoros rutinok idegrendszeri reprezentációjának : megerősödése átrendeződése
Hosszú távú tanulás - zenészek Schlaug. et al. (2005) Ann. NY Ac.ad. Sci.
Hosszú távú tanulás - zenészek Schlaug. et al. (2005) Ann. NY Ac.ad. Sci.
Hosszú távú átrendeződés Vakok Braille olvasásakor aktiválódik a látókéreg Sadato et al. (1996) Nature
Hosszú távú átrendeződés A vakok látókérgének funkciója Amedi et al. (2003) Nature Neurosci