HOMEWORK #2:

Checklist:
  (a) due Sat 3/24/06, under my door (CSB 171)
  (b) use this page as the first page
  (c) staple all pages together including the 'slightly longer question'
  (d) one page maximum per answer (less than a page is better!)
  (e) each answer on a separate page (front side only)
  (f) figures on same page as answer
  (g) word-processor text with hand-drawn figures preferred
  (h) list the names in your working group (if any besides you)


1. (a) Explain why there is an "aperture problem for pattern translation", using evidence from cell responses in layer 4B of primate V1 and area MT. (b) Outline a mechanism by which this problem could be solved, mentioning both direction and speed. (c) Third, describe a higher order 'aperture problem' that is solved by some neurons in area MSTd, including what can't MT see, why can't it see it, and how come MSTd can see it. In the process, give an informal description of a model for how MSTd neurons might achieve approximate position invariance in their responses to particular flow field patterns (MSTd model in reader). (d) Finally, say briefly why it makes sense to describe color constancy across different illumination conditions as an 'aperture problem'.

2. (a) Summarize the evidence for viewer-centered representations of visual objects in primate visual inferotemporal cortex presented Logothetis and Pauls, Wang et al., and Rolls et al. in the readings. (b) What evidence is there for (and/or against) translation invariant responses in inferotemporal cortex.

3. (a) Make a diagram of the dorsal column and spinothalamic pathways to somatosensory cortex, and include known somatosensory cortical areas in primates, distinguishing only "hand", "foot", and "face" (different sources will disagree). (b) Describe at least 4 experiments suggesting that correlated activity dynamically maintains cortical somatosensory maps. (c) What basic dichotomy in the time course of the response of different somatosensory receptors is shared with ganglion cell types in the visual pathway? (d) Why do some muscle spindles need their own muscle-spindle-muscles? (e) Describe two situations for a muscle: where the muscle spindles fire strongly and Golgi tendon organ responses are weak, and another situation where the opposite is true.

4. The cochleas transduce sound into neuronal firing patterns distributed across the fibers of the two auditory nerves. (a) Describe the primary stages by which interaural time difference (ITD) is calculated in the owl auditory system using these two sets of signals and how the phase ambiguity problem is solved. Consider the signals found in the magnocellular cochlear nucleus (NM), nucleus laminaris (NL), the central nucleus of the inferior colliculus, lateral subsection (ICc lat), and the external nucleus of the inferior colliculus (ICx). Be sure to define the phase ambiguity problem at the neural level. (b) Animals with small heads have small delays between the ears. What problem does this pose for an owl-style concidence detection and ambiguity resolution mechanism like the one you just outlined?

5. (a) What difference between the bat and primate auditory system creates the bat acoustic fovea? (b) If a moth detects a bat call and begins to mistakenly fly toward from the attentive bat, will the bat increase or decrease the pitch of its voice? (and why does it do this?) (c) Explain why vowels sound almost the same when the pitch of a person's voice is raised or lowered. (d) What aspect of bat CF/CF processing resembles vowel recognition across different human speakers? (e) What are "formant transitions", what causes them, and how are they similar and different from the frequency-modulated parts of bat calls?

6. The superior colliculus is often described as having a superficial retinal map overlying an intermediate saliency map overlying a deeper motor (saccade) map. Ordinarily, localized activation of the superficial layers elicited by a target stimulus is followed by activity in the corresponding region of the underlying motor map, which generates a saccade that moves the center-of-gaze onto the target. (a) Describe a sequence of fixations where the locus of activity in the superficial layers caused by a target eventually leads to activity at a different locus in the intermediate and deep layers of the colliculus (leading to a saccade to that target). (b) Suggest a plausible neural pathway (nuclei, connections) that could be causing these shifts of activity in the colliculus. (c) Describe the problem that off-center gaze creates for integration of auditory and visual cues for target location, and say how the superior colliculus solves it (Jay and Sparks expt on auditory superior colliculus). (d) Describe the evidence for head-centered visual coordinate frames in VIP (Duhamel, Colby).

7. We discussed cells in two groups of nuclei in the limbic system: (i) the CA fields of the dorsal/posterior hippocampus and (ii) the postsubiculum, mammillary nuclei, and anterior (LD) thalamic nuclei. These two types of cells use environmental cues and other information about internal brain states to calculate aspects of the animal's current position. Briefly explain the following. (a) Why are the cues called "distal"? (b) What feature of the animal's position do neurons in these two regions signal? (c) Why are these cells not thought to be merely cue-driven? (d) What additional information beyond the distal cue information must be being used to drive the two kinds of cells' responses? (be as specific as you can)

8. A long series of experiments (R. Thompson lab) suggested that the cerebellum is involved in storing the learned sensorimotor mapping in a simple conditioning task. (a) First, descibe the task, (b) then review evidence that conditioned response (CR) requires the cerebellum, and (c) that the conditioned stimulus (CS) is processed by cerebellum, and (d) describe why the results were surprising (beyond just 'it turned out to involve the cerebelllum'...). Finally, (e) starting at the sensory neocortex of a mammal, list the structures (in correct order!) that information has to pass through on its way to the spinal cord via the cerebellum.

9. The striatum is in a position to influence motor behavior. (a) Make a diagram of the major connections between the caudate/putamen, globus pallidus, substantia nigra, subthalamic nucleus, superior colliculus, and motor cortex. See if you can find information about the connections of the three parts of the globus pallidus, the two parts of the substantia nigra, and the two parts of the subthalamic nuclei out of the reader and/or the literature (N.B.: there remains some debate about these connections). (b) Medium spiny cells in the striatum (the main inhibitory output cells of caudate/putamen) exhibit so-called "up" and "down" states. It has also been discovered that these cells have an "anomalous rectifier" potassium current that is turned off by depolarization. Why is this current called "anomalous" and in which state ("up" or "down") might you expect this channel to conduct potassium?

Slightly Longer Question (choose one--2 to 4 pages, attached)

You should discuss experiments from more than one of the major fields of study mentioned in class (these include [at least]: EEG/MEG, fMRI/PET/optical, human brain lesions, animal brain lesions, single-unit neurophysiology, neuroanatomy, neural network models, human behavior, animal behavior). If you have a favorite question that you would like to answer instead, state that question at the top of the paper on a separate line, and then follow the guideline above in answering it. Obviously, it should relate reasonably closely to a topic covered in the course.

1. Discus neuroanatomical, neurophysiological, and neuroimaging evidence for mechansims and stages of complex motion processing (beyond local translation).

2. Discuss neurophysiological and computational models of selective visual attention.

3. Discuss neurophysiological data and computational models of sleep.

4. Compare birdsong and human speech sounds with regard to the spectra of individual segments, segment sequences, development, and function, briefly mentioning behavioral, neurophysiological, and neuroanatomical data.

5. The cerebellum has long been a topic for neural modelers (e.g., Marr). Cerebellar experiments (e.g., Shambes, Bower, Llinas, Thompson), however, have revised earlier ideas about cerebellar function. Outline the anatomical organization of the cerebellum and then discuss recent evidence for how the cerebellum is involved in sensorimotor learning.

6. Consider recent evidence for how the organization and connections of visual cortical areas are specified in development, and comment on the relation between intracortical, extracortical (e.g., thalamic), and environmental factors.

7. Discuss several main classes of 'neural' network models, the features of real networks upon which they are based, how they differ in fundamental ways from the real networks, and finally, outline some of the advantages and disadvantages of neural realism.

8. Describe evidence for functional segregation of sensory input into parallel streams in several different sensory modalities (visual, auditory, somatosensory, electric).

9. Compare and contrast animals and humans with respect to the behavioral effects of hippocampal damage as well as the results of recording experiments in the hippocampus (there are, in fact, a number of hippocampal recording experiments in humans).

10. Describe the standard model for the generation of saccades by the superior colliculus and brainstem reticular eye movement control structures (D.R. Robinson). Then discuss recent evidence for shifting maps in the colliculus and parietal cortex (e.g., Sparks, Guitton, Colby) suggesting that eye position information is used to rapidly update a retinotopic map of visual targets.

11. Language deficits after cortical lesions have typically been discussed in terms of breakdowns in phonology, perception of word forms, writing, syntax, semantics, and pragmatics. Studies in primate cortex, on the other hand, have revealed a mosaic of visual, auditory, somatosensory, motor and limbic areas. Discuss at least one lesion-induced language deficit and related fMRI language experiments (from the huge horde in Neuroimage) from these two somewhat incommensurable perspectives.

12. It has been suggested that the striatum is involved in the hierarchical control of motor behavior. Behavior takes place over a range of time scales -- for example, a day long goal of finding food can be divided up into searches of various possible food stores, which can in turn be divided up into parking, walking in, checking the vegetable isle, which can be divided up into individual steps of the left and right feet and grasping movements. The basal ganglia (caudate, putamen, globus pallidus, substantia inominata, ventral pallidum) seem to be involved at all these time scales. Briefly review anatomical data about multiple loops through striatum that might support these hierarchies (Strick, Graybiel).