Psychophysical Methods
Hearing
Taste
Olfaction
Vision
Speech

Psychophysical Methods

Channel capacity

Coren, Ward, and Enns (1993) defined channel capacity as "the limit to the number of bits of information an observer can transmit on a single stimulus dimension." Channel capacity represents an index that defines how many bits of information an observer can transmit perfectly. Or more specifically, channel capacity of the perceiver is a function of the number of stimuli that can be accurately be discriminated.

The maximum amount of discriminable information in absolute judgments varies from 3 to 12 and varies as a function of several experimental factors: (1) the physical range of stimulus variation; (2) the spacing of the stimuli within a prescribed range; (3) the number of stimuli and response categories; and (4) an observer's opportunity to obtain knowledge of the results concerning judgment accuracy. The number of channel capacity in bits has come to be termed Miller's "magic number of seven plus or minus two" (Dember & Warm, 1979).

Merkel (1885) attempted one of the first experiments on stimulus uncertainty on choice reaction time (RT) by manipulating arabic and roman numerals stimuli varying in number from 1 to 10 and having subjects respond to the experimental task by pressing one of ten keys. Merkel found that RT increased in proportion to the logarithm to the base 10 of the number of alternatives available. Hicks (1952) replicated Merkel's findings by manipulating one to ten lamps arranged in an irregular circle where the subject responded to the illumination of the lamps by pressing one of ten Morse Code keys. He found that RT increased as a linear function of the log₂ of the number of stimulus alternative available or, in other words, RT was proportional to stimulus uncertainty (Dember & Warm, 1979). This finding, therefore, elucidates that choice RT is a linear function of the amount of information available by the stimulus, and choice RT between conditions can only be compared when the number of response alternatives are identical (Coren et al., 1993).

The just noticeable difference, or jnd, is a measure of how different the stimuli can be from each other and have the difference be just noticeable. The difference of 1 jnd is the same as the difference of 1 difference limen or dl (a measure also used in discrimination tasks) and are, therefore, used interchangeably. Therefore, as Coren et al. (1993) defined it, the jnd is "the subjective experience of the difference threshold; the sensation difference between two stimuli separated by one difference threshold." A stimulus at absolute sensory threshold generates 0 units of sensory magnitude on Fechner's scale whereas a stimuli that generated stimulus intensity 1 difference threshold above absolute threshold had 1 unit of sensory magnitude or 1 jnd. The jnd is also used in Weber's Law which concerns the relationship between the jnd and the standard stimulus:

jnd/S = k (cited in Dember & Warm, 1979)

Our perceptual experience includes the register of sensory properties of stimuli that do not remain constant but vary across a considerable range of magnitudes. Light, for example, can range from very bright to very dim, sound can vary from very loud to very soft, and so forth. The question then is, how can psychological measures of physical event be coincided with physical measures? Such a question is important because the two, that is the perceived and physical registration of the magnitude of a stimuli, are not the same. If a light source is increase by twice as much in intensity, is the perceived brightness the same? It is not. Therefore, the perceptual system is not a simple, linear reflector of variations in the physical values of stimuli and, to account for this, means have been developed to measure this and establish a link between perceived events and the physical world. Measurement refers to the "assignment of numbers to objects or events according to rules." An experimenter needs to attach numbers to quantify the behavior of subjects in an experimental design which is usually defined by the operational definition of the construct being measured. An operational definition concretely explicates what must be done to measure that construct. For example, in psychophysics, detection, identification, discrimination, and scaling represents the concerns of perceptual psychology and the use of the four is determined by the research question and subsequently the operational definition assigned to the construct. However, not all variables can be measured quantitatively and, therefore, different scales have been developed to measure different types of constructs. The four scales are the nominal, ordinal, interval, and ratio scales.

Weber's fraction

"A measure of the overall sensitivity of a sensory system to differences along a stimulus continuum" (Coren et al., 1993). Weber's law is a measure of differential sensitivity where the smaller an increment in energy necessary for a difference to be perceived by the observer, the greater the sensitivity of the observer's sensory system. Weber's Law can be stated as ^I/I = k or as jnd/S = k where I is the magnitude or intensity of the standard stimulus, ^I is the threshold increment, or the jnd, and K is a constant or the Weber Fraction. For illustration of Weber's Law, suppose a 200-gram weight is utilized as a standard and it is found that 4 grams is needed to obtain a jnd. If a 1000-gram weight is then used, 20 grams in now need to obtain the jnd. Therefore, the Weber fraction provides an index to the discriminatory power of the sensory modality where the smaller the Weber fraction, the more sensitive the discrimination. In addition to providing a measure of an observer's sensitivity to stimuli differences, the fraction potentially can be utilized as a prediction tool. However, validity studies of the Weber fraction have shown it to valid in the middle ranges of stimulation with deviations from the expected constancy at the extremes (Dember & Warm, 1979).

Error of anticipation

In the method of limits, inequalities exist between ascending and descending trials because of two response tendencies evident in the response patterns of subjects: (1) errors of habituation and (2) errors of anticipation. Within any series of judgments, subjects become aware that their responses shift from one category to the other. Error of anticipation becomes evident when subjects operate from the principle that "the stimulus is likely to be different from last time, so I'll change my answer." As a result, they "jump the gun." When the ascending averages (RL) is smaller than the descending RL the error of anticipation may be greater than the error of habituation. Subjects claim they detect the stimulus when it is not there in ascending trials; on descending trials, they claim that they can no longer detect the stimulus when it is in fact there. Errors of anticipation can be dealt through the use of alternating ascending and descending trials. Also, if we can assume that subjects make as many errors of habituation as errors of anticipation, the errors will cancel each other out. Finally, thresholds may represent a subject's systematic tendency to supply the same response or a different response each time, therefore, the same number of ascending and descending trials should be included and the starting point should be systematically varied (Dember & Warm, 1979).

A measure of sensitivity in Signal Detection Theory that is utilized in place of d' when the assumption of normality of the distribution is violated and, therefore, a nonparametric statistic is needed. The equation for A' is : 1 - 1/4 [p(fa)/p(hit) + 1-p(hit)/1-p(fa)]. A' ranges from .5 to 1 where A'=.5 is chance level equal to a d' of 0 and an A'= 1 is ideal performance equal to a d' of infinity.

Sone

Some evidence has shown that a lack of equality of jnd's along prothetic continua exists using procedures where perceptual magnitudes are measured by using direct estimate, rather than Fechner's indirect approach. These "ratio-scaling" procedures state that observers can validly assign numerical values to the perceptual experience and also that if two stimuli are presented, a perceiver can assign a number to the ratio of the perceptual magnitude experienced. Thus, an observer can accurately state that one stimulus is twice that of another stimulus or, in other words, the second stimulus has 1/2 the magnitude of the first. One of the principal ratio-scaling techniques is the "fractionation" method that asks observers to set the physical value of a stimulus to be half as intense as another stimulus. Stevens (1936) used this method to construct the sone scale (Dember & Warm, 1979). The sone scale "is a scale used to measure the loudness of a sound; also the unit of that scale -- a tone of 1000 Hz at 40 dB has a loudness of 1 sone" (Coren et al., 1993).

Residual stimuli

Helson (1964) stated that an organism attempts to accommodate itself to the changing environment that surrounds it (cited in Coren et al., 1993). This accommodation involves the establishment of a reference level against which all other stimuli are judged against as a standard. Stimuli below this referent level, or adaptation level, are thought to be weak in intensity whereas those stimuli above the adaptation level are perceived to be strong in intensity and stimuli at or around the adaptation level are thought to be neutral. Therefore, stimuli intensity is a relative perception based upon subjective adaptation level that are govern by different types of effects from classes of stimuli. One class of stimuli is residual stimuli where the past experience of the observer influence the perception of stimuli. An example was provided in the book of the case where sportscasters interview both basketball players and jockeys. The perception of the height is a weighted combination of the focal stimuli, or the physical height of the sportscaster, the background stimuli, the height of the surrounding athletes, and the residual stimuli, or the heights of all persons that have been previously encountered. Therefore, residual stimuli are those stimuli that are not longer present but still affect the perceptual process by influencing the current adaptation level.

Information transmission

Coren, Ward, and Enns (1993) defined channel capacity as "the limit to the number of bits of information an observer can transmit on a single stimulus dimension." Channel capacity represents an index that defines how many bits of information an observer can transmit perfectly.

The amount of information transmitted is usually depected in a schematic illustration that represents the amount of information that can be transmitted in an absolute-judgment task as a function of the amount of input information that is provided and is plotted against a ideal or theoretically perfect accurate performance. The figure illustrates that the extent to which errors are introduced in the judgments, decrements in the amount of information transmitted become more evident. Information transmission is also often depicted in a confusion matrix which tables the number of responses made to a given stimulus.

Fechner built on Weber's Law which states that, though the absolute value of a jnd increases as stimulus magnitude increases, the relative value of the jnd remains constant. Fechner stated that a subject rating scale could be utilized that had an arbitrary referent, or starting, point for absolute threshold measurement and use the jnd as an "indirect" but constant unit of measurement (Dember & Warm, 1979). Essentially, he constructed a scale by adding the jnds together. In order to determine the relative strength of the stimuli, a summation of the jnds that fell between the stimuli and the absolute threshold was needed where, for example, if the value of the stimulus was 10 jnds above the absolute threshold, the perceived magnitude was 10.

(1) Weber's law was true

(2) the perceived magnitude of the stimulus is equal to the number of jnds that had preceded it on the scale

(3) the jnds are each subjectively equal and the difference between two jnds would remain constant across the magnitude scale (Dember & Warm, 1979).

Assuming that Weber's law was indeed correct, he could mathematically alter the Fundamentalformel, or fundamental theory of Weber, from: ^R=k (^S/S) into what he termed Massformel (measurement formula) or what is now known as Fechner's law: R=k log S where k is a constant that is dependent on the sensory modality, the units of measurement used, and the zero point that is chosen and S is the physical value of the stimuli. Fechner's equation states that the perceived magnitude increases as a logarithmic function of the stimuli that is being perceived by the observer. What is especially characteristic of this equation is that it takes successively larger and larger physical values of the stimuli for equal increments in the perceived magnitude. Also, by plotting perceived magnitude in arbitrary units as a function of the log of the stimulus, the psychophysical relation between physical magnitude and perceived magnitude changes from a curvature relationship to one of a straight line, linear relationship as shown in the figure below (Dember & Warm, 1979).

Disadvantages of Fechner's Law.  Fechner's Law can be attacked on several grounds. First, Weber's law generally does not remain valid for extreme ends of the sensory continuum and, therefore, Fechner's summation of the jnds may introduce measurement error. Secondly, the use of the absolute threshold as a starting point may not be judicious because the measurement of the absolute threshold, as Dember and Warm (1979) state, is "itself open to serious question." Finally, the assumption that all the jnds are subjectively equal in the magnitude scale continuum has not been supported by research.

Stevens (1957) was one of the first to really elucidate these problems especially the last one. Stevens stated that two broad kinds of perceptual continua exists.
 

Prothetic stimuli are concerned with "how much" and are stimuli in which discriminations between them are made by an additive process on the physiological level such as loudness, heaviness, etc.

Methatic continua are concerned with "what kind" where discriminations are made through a subtractive process at the physiological level such as pitch, azimuth, and apparent inclination.
 

Fechner's law has shown to generally hold for metathetic dimensions, but it does not hold for prothetic stimuli as has been shown by "ratio-scaling" studies where it has been shown that a lack of equality of jnd's along prothetic continua exists using procedures where perceptual magnitudes are measured by using direct estimate, rather than Fechner's indirect approach. These "ratio-scaling" procedures state that observers can validly assign numerical values to the perceptual experience and also that if two stimuli are presented, a perceiver can assign a number to the ratio of the perceptual magnitude experienced. Thus, an observer can accurately state that one stimulus is twice that of another stimulus or, in other words, the second stimulus has 1/2 the magnitude of the first.

Steven's Law.  Stevens developed one type of ratio-scaling technique called magnitude estimation where observers assign numerical values to stimuli in proportion to the perceived subjective magnitudes. Another technique developed by Stevens is Magnitude Production which is the inverse of magnitude estimation where the experimenter presents stimuli and asks the subject to adjust the magnitude to subjective values. By using these technique, Stevens indeed found that psychological magnitude did not increase as a log function of stimulus magnitude as predicted by Fechner's law, but instead was a power relationship where the perceived magnitude is proportional to physical magnitude raise to some power: R = kSⁿ where R is the psychological response to the stimulus (S) and k is a constant the depends on the physical stimulus. S is raised to some n power based on the physical stimulus being studied. Since the equation involves raising some stimuli to a certain power, the equation is referred to as Steven's Power Function. This equation states that as stimulus magnitude increases geometrically so also does perceived magnitude such that equal stimulus ratios produce equal perceptual ratios. This power function has provided greater variation in psychological scales because power functions differ across different sensory modality as is illustrated in the table below (Dember & Warm, 1979).

Understanding that subject's perceive magnitude rating may be biased through a moderately sophisticated knowledge of the use of numbers, Stevens argued that the validity of his finding would be enhanced if it could be replicated in an environment that did not rely on the use of numbers to rate perceived magnitude of a stimulus. Stevens utilized cross-modality matching where the subject is asked to judge stimuli in one mode of perception such as hearing by providing estimations based on another sensory modality such as sight. As the book pointed out, a perceiver may be asked to squeeze a handgrip until the pressure felt as strong as a light source being observed. If the data obtained from such cross-modality matching experiment are plotted on a log-log axes, the data fall on a straight line and, therefore, still follow the power function for sensation intensities. Cross-modality matching is often more difficult for the subject to make magnitude estimations than by employing the direct magnitude estimation technique because the subject must adjust one sensory continua to make a response on an independent sensory continua in lieu of simply providing a numerical value. New refinements have since made the process a bit easier. One modification to the cross-modality matching technique is mixed-modality scaling where the observers judge two different sets of sensory stimuli where both sensory stimuli are intermixed in the experiment. Therefore, it differs from cross-modality matching in that subjects do not actually match sensory magnitudes. Instead, the perceiver tries to utilize the same scale for both stimuli as they are alternated in the experiment.

Category scaling and multidimensional scaling

Coren et al. (1993) defined category scaling as "a psychophysical scaling method in which stimuli are grouped in a predetermined number of categories on the basis of their perceived intensity." Subjects are provided with a number of categories in which the subject must make a judgment and place that stimuli in one of the categories based on what he or she believes is the weight, perceived brightness, etc. An example that was provided by the book involves an experiment by Sanford (1898) that provided observers with a number of envelopes in which they were to categorize the envelopes into five categories based on their respective weights. The difference between the categories was argued to be the same in relative distance such that the difference in weight between categories 1 and 2 would be the same as the difference between categories 4 and 5. What is characteristic of this method is that there are no right or wrong answers because, as Coren et al. (1993) point out, the very nature of the experiment prevent such measures of successful and unsuccessful performance in that there is no way to know in advance what is the right categorical judgment. The data are averaged and are plotted as the average category label as the Y axis and Stimulus intensity as the X axis. The resulting curve is a concave downward curve that closely approximates Fechner's Law.

Multidimensional scaling bears some semblance to category scaling in that both represent data in physical units in relation to how similar or dissimilar the data are relative to other data and both measure these on an interval scale. In contrast to category scaling, however, multidimensional scaling does not use categories but instead a set of stimuli are presented to the subject in which the subject must arrange them in such a fashion as to represent them on a continuum. A similarity matrix is formed by these arrangements where pairs of stimuli that appear to be more similar receive high similarity ratings. The rating are then plotted on a spatial map where like data points on the multidimensional map are place closely together in space and dissimilar stimuli are placed farther apart. Multidimensional scaling was developed by Torgerson and Shepard who argued that psychological similarity could be represented physically as distances on a two-dimension plane or spatial map. An example of a multidimensional map can be seen below.

However, the two are remarkably similar in that the subject is allowed freedom in providing magnitude estimations and, therefore, are not as hampered by other psychophysical methods that impose more structure on the methods the subject uses for response ratings. However, each scaling technique provides the subject with a set number of response criteria (e.g., 1-10 for multidimensional scaling) and, therefore, do limit the precision as to the degree of perceived magnitude. One advantage over category scaling that multidimensional scaling retains is that psychological attributes that do not coincide with physical dimensions can still be represented in a spatial map. For example, a study by Shepard and Chipman (1970) examined the shape of states and asked subjects to make rating based on the correspondence between people's judgments of physical shapes and their judgments of mental shapes. The results of the multidimensional scaling analysis was plotted on a diagram in which states judged similar to each other were located near each other in the diagram (Dember & Warm, 1979). Such a depiction of the metathetic data where the diagram is meaningful and significant in illustrating functional relationships could not have been achieved by making an X-Y plot as required by category scaling.

Response continuum and metathetic continuum
 

There exists three main types of continuum:

The physical continuum represents real energy that exists in the physical world. It is comprised on the entire range of electromagnetic spectrum of energy.

However, as perceptual observers, the entire physical continuum is not accessible to humans. The range to which we are sensitive to the physical continuum is the subjective continuum.

The third continuum is represented by the response continuum which is the range to which subjects say the detect the physical stimuli. The response continuum, of which all psychophysics is based upon, lies just inside the subjective continuum where the extreme ends of the continuum become a probable detection of threshold arising from such measurement errors as faulty equipment, uncooperative subjects, or the body itself interacting with the stimuli to produce a response (e.g., 20% of neurons may be firing at any given moment). In measuring the response continuum, the question arises as to how different stimuli are to be measured as all sensory qualities cannot be scaled in the same way as others because, as the book points out, some perceptual experiences have an underlying aspect of intensity, such as brightness, whereas others, such as color, do not.

Prothetic stimuli are concerned with "how much" and are stimuli in which discriminations between them are made by an additive process on the physiological level such as loudness, heaviness, etc.

Methatic continua are concerned with "what kind" where discriminations are made through a subtractive process at the physiological level such as pitch, azimuth, and apparent inclination.

Each just noticeable differences contributes equally to perceived magnitude across the sensory scale only with metathetic stimuli but not with prothetic stimuli (Dember & Warm, 1979). Therefore, the demarcation between response continua illustrates that perceptual experiences are governed by different laws as a function to which type of stimuli being observed. In order to measure the response continuum, the type of continua must be taken into account as the metathetic continuum can only be measured on a nominal scale. Therefore, to compare and contrast the response continuum and the metathetic continuum, the response continuum represents the sensitivity of the observer to stimuli that are either of prothetic and metathetic stimuli. The response is governed as to the question that arises from the perceptual experience; that is, whether it is "how much?" or "what kind?" determines how we perceive the stimulus.

The visual system is very complex and those interested to find out more information are directed to Sekular, R. & Blake, R. (1994). Perception (3rd Edition). New York: McGraw-Hill for more information.  However, for the purposes of the human factors course, the information below should suffice for most purposes.
 

THE ORBIT

The orbit contains the eyeball, extrinsic eye muscles, lacrimal gland, cranial nerves and blood vessels. The roof, floor and medial and lateral walls of the orbit are made up of seven bones. These bones include: frontal (roof, medial wall and supra-orbital margin); maxillary (floor, medial wall, infra-orbital and medial orbital margins); lacrimal (orbital rim and medial wall); ethmoid (medial wall); sphenoid (roof,lateral and medial walls); palatine (floor); zygomatic (lateral wall, floor and lateral orbital and infra-orbital margins).

There are six foramina within the orbit that transmit important structures for the eye and its accessory structures.

1) The supra-orbital foramen is located in the frontal bone on the supra-orbital ridge. It transmits the supra-orbital nerve that innervates the eyebrow, eyelid and frontal sinus, and also transmits the supra-orbital vessels which also supply these areas.

2)  The nasolacrimal canal is in the lacrimal bone and it transmits the tear duct that drains into the nasal cavity.

3) The optic canal is located in the posterior part of the roof, through the sphenoid bone and transmits the optic nerve (CNII) as well as the ophthalmic artery.

4) The superior orbital fissure is also located in the sphenoid bone and it transmits the oculomotor nerve (CNIII), trochlear nerve (CNIV), abducens nerve (CNVI), ophthalmic nerve (CNVI) and the ophthalmic vein.

5) The inferior orbital fissure is found at the sphenoid and maxillary bones. It transmits the maxillary branch of the trigeminal nerve (CNVII).

6) The infra-orbital foramen is located in the maxillary bone at the medial end of the infra-orbital margin. It transmits the infra-orbital nerve and artery.

ACCESSORY STRUCTURES OF THE EYE

The Eyelids / Palpebrae

The eyelids are continuations of skin that serve to lubricate the eye and keep dust and debris away by continually blinking. They are strengthened by the tarsus or tarsal plate, which is a dense connective tissue band. The firbres of the orbicularis oculi are located in the connective tissue between the tarsus and the skin. Tarsal glands are embedded in the tarsus, of both the upper and lower eyelids, which produce a fatty secretion that lubricates the eyelids and prevent them from sticking together.

The palpebral fissure separates the free margins of the upper and lower eyelids and the medial canthus and lateral canthus connect the upper and lower eyelids at their medial and lateral ends respectively.

The eyelashes prevent particles contacting the eye and are associated with the glands of Zeis, which are large sebaceous glands. The Meibomian glands are modified sebaceous glands that lie along the inner margin of the eyelid and serve to keep the eyelids from sticking to one another.

The lacrimal curuncle is a soft mass of tissue located at the medial canthus and it contains glands that produce thick secretions.

The conjunctiva of the eye is the stratified squamous epithelium that covers the inner surface of the lids and is continuous over the outer surface of the eye. It is divided into the palpebral conjunctiva, which covers the inner surfaces of the eyelids, and the bulbar conjunctiva, which covers the outer surface of the eye and extends to the edges of the cornea.

The Lacrimal Apparatus

The roles of the lacrimal apparatus include producing, distributing and removing tears from the eye. It consists of the lacrimal gland, the lacrimal canals and the nasolacrimal duct.

The lacrimal gland lies in the depression of the frontal bone inside the orbit, superior and lateral to the eyeball. It produces most of the volume of tears, which is watery and alkaline and contains lysozyme. About 1ml/day is produced and mixed with the secretions from the sebaceous glands, to assist lubrication and prevent drying out of the eye.

The action of blinking sweeps tears across the surface of the eye towards the medial canthus, where they accumulate at the lacrimal lake overlying the lacrimal caruncle. The lacrimal lake subsequently drains through two lacrimal puncta (small pores), which in turn empty into the lacrimal canals. The canals then drain into the lacrimal sac within the lacrimal groove of the orbit.

From the lacrimal sac, the tears drain through the nasolacrimal canal and enter into the inferior meatus of the nasal cavity.
 
 

THE STRUCTURE OF THE EYE

The eye contains three layers or tunics called the outer fibrous tunic, the intermediate vascular tunic and the inner neural tunic.

The eyeball itself is hollow and is divided into two cavities by the ciliary body and the lens of the eye. These are the large posterior cavity, or vitreous chamber, and the small anterior cavity.

The anterior chamber of the anterior cavity extends from the cornea to the iris, and its posterior chamber lies between the iris and ciliary body and lens. Both of these chambers contain aqueous humor, which is a fluid that circulates within the anterior cavity. The aqueous humor is produced by active secretion from the epithelial cells of the ciliary body's ciliary processes and provides a fluid cushion for the eye as well as a route for nutrient and waste transport.

The posterior cavity is filled with a gelatinous substance called vitreous humor. It contains collagen fibres and proteoglycans which are produced by specialised cells in the vitreous body. This substance helps maintain the shape of the eye and supports the retina.

 
Fibrous Tunic

The fibrous tunic consists of the sclera and cornea and provides mechanical support and physical protection, attachments for the extrinsic eye muscles and contains structures which assist the eye to focus during vision.

The sclera is the "white of the eye" and consists of dense, fibrous connective tissue that covers the anterior and posterior surfaces of the eye. The extrinsic muscles attach to it by way of binding their muscle fibres with the collagen fibres of the sclera. The surface of the sclera also contains blood vessels that penetrate it to reach the internal structures of the eye.

The cornea is transparent and structurally continuous with the sclera at the limbus (border between the sclera and cornea). It has corneal epithelium which is squamous and consists mainly of a dense matrix with multiple layers of collagen fibres.Whilst there are no blood vessels in the cornea, there are many nerve endings which are quite sensitive.

The lens lies posterior to the cornea and is attached to the ciliary body by suspensory ligaments. Its role is to focus the visual image on the retina by changing its shape. It is primarily composed of concentric layers of cells with a dense fibrous capsule covering its entire surface. These cells are highly specialised with no nuclei or other organelles and are filled with transparent proteins called crystallines.
 

Vascular Tunic

The vascular tunic includes the iris, ciliary body, choroid and the intrinsic eye muscles and contains blood vessels and lymphatics. Its functions include, providing a route for blood vessels and lymphatics that supply the tissues of the eye, regulating the amount of light that enters the eye, secreting and absorbing the aqueous humor within the eye and controlling the shape of the eye to aid in the focussing process.

The iris can be seen through the cornea and contains blood vessels, pigment cells and two layers of smooth muscle fibres. The central opening of the iris is the pupil and the size of this opening is controlled by the two layers of smooth muscle. The pupillary constrictor muscles act like a sphincter and are located in concentric circles around the pupil. The pupillary dilator muscles radiate away from the pupil. Both of these layers are innervated by the autonomic nervous system.

The body of the iris contains vascular, pigmented, loose connective tissue. The anterior surface has no epithelium, but an incomplete layer of fibroblasts and melanocytes. Its posterior surface has pigmented epithelium and melanin granules and is strictly part of the neural tunic. The melanocytes are scattered throughout the body of the iris and their density and distribution contribute to the colour of one's eyes. The iris attaches, at its periphery, to the anterior portion of the ciliary body.

The ciliary body begins at the junction between the cornea and the sclera and extends posteriorly to the anterior serrated edge of the neural retina (ora serrata). It consists of the ciliary muscle, a smooth muscle ring which passes into the interior of the eye. The epithelium of the ciliary body forms folds called ciliary processes, into which the suspensory ligaments of the lens attaches.

The choroid is a vascular layer that separates the fibrous and neural tunics, posterior to the ora serrata. It is covered by the sclera and attached to the outermost layer of the retina. Its main constituents are a capillary network that transmits oxygen and nutrients to the retina and melanocytes, which are mostly found near the sclera.
 

Neural Tunic

The neural tunic, or the retina, is the innermost layer of the eye. It has an outer pigmented layer, which is responsible for absorbing light that passes through the eye, and a thick inner layer called the neural retina, which contains the photoreceptors that respond to light, supporting cells and neurons for preliminary processing and integration and blood vessels that supply tissues in the posterior cavity of the eye.

The retina is divided into several layers of cells types. The layer closest to the outer pigmented layer contains the photoreceptors (rods and cones). The rods are responsible for detecting the intensity of light and enable us to see in dim light. The cones are responsible for colour vision and are subdivided into three types. Various combinations of stimulation of these three types of cones provide us with colour vision.

The macula lutea is the site on the retina where there are no rods and where the visual image arrives after passing through the cornea and lens. The fovea is the centre of the macula lutea that has the highest concentration of cones and provides us with the sharpest vision.

The next layer of the retina contains the bipolar cells which synapse with the rods and cones. The bipolar cells in turn synapse with the next layer, consisting of ganglion cells.

There are numerous interneurons that can modify the synaptic transmission, called horizontal cells. These are located between the photoreceptors and bipolar cells. A similar layer exists between the bipolar cells and ganglion cells and are called amacrine cells. Both of these cell types can modify the sensitivity of the retina. The interplexiform cells are another type of association neuron which are dispersed among the cell bodies of the bipolar cells. They provide a feedback loop as they are postsynaptic to the amacrine cells and presynaptic to the horizontal and bipolar cells.

The cells of Muller are radial neuroglial cells which extend from the nerve fibre layer of the retina and the vitreous body, to the junction of the inner segments of the photoreceptors and their fibres. These cells are mostly supportive cells in the retina.

The optic disc is a circular region medial to the fovea where axons from about one million ganglion cells converge. It is also the site of origin of the optic nerve. The central retinal artery and vein which supply the retina also emerge on the surface of the optic disc. This area contains no photoreceptors and therefore light arriving at this point goes unnoticed and is also known as the blind spot.

The Histological Layers of the Retina

Haematoxylin and eosin stained sections of the retina, reveal ten layers in relation to the cells that constitute the retina.

Layer 1: pigmented epithelial cells

Layer 2: the outer and inner segments of rods and cones

Layer 3: the outer limiting membrane (where the outer ends of the Muller's cells contact the photoreceptor cells)

Layer 4: the outer nuclear layer (consisting of the nuclei of the rods and cones)

Layer 5: the outer plexiform layer (consisting of photoreceptor fibres and bipolar cell dendrites)

Layer 6: the inner nuclear layer (consisting of the nuclei of bipolar, horizontal, amacrine, interplexiform and Muller's cells)

Layer 7: the inner plexiform layer (consisting of presynaptic dendrites of bipolar cells & postsynaptic dendrites of ganglion
cells)

Layer 8: the ganglion cell layer (cell bodies of ganglion cells)

Layer 9: the nerve fibre layer (axons of ganglion cells)

Layer 10: the inner limiting membrane (consiting of the expanded ends of Muller's cells)
 
 

MOVEMENTS OF THE EYE

The gross movements of the eyeball result from the action of the six extrinsic eye muscles.

The inferior rectus muscle acts to move the eye in a downward direction. Its attachments are the sphenoid bone around the optic canal near the junction of the inferior and superior orbital fissures (via the common tendinous ring) and the infero-medial surface of the eyeball.

The medial rectus muscle adducts the eye. Its attachments are the sphenoid bone (via the common tendinous ring) and the medial surface of the eyeball.

The superior rectus muscle elevates, adducts and medially rotates the eye. Its attachments are the sphenoid bone (via the common tendinous ring) and the superior surface of the eyeball.

The inferior oblique muscle elevates the medially rotated eye, abducts and laterally rotates the eye. It attaches to the maxillary bone in the floor of the orbit and the infero-lateral surface of the eyeball.

The superior oblique muscle depresses the medially rotated eye, abducts and medially rotates the eye. It attaches to the sphenoid bone around the optic canal and the supero-lateral surface of the eyeball.

The lateral rectus muscle abducts the eye. It too is attached to the sphenoid bone (via the common tendinous ring) and the lateral surface of the eyeball.

The oculomotor nerve (CNIII) innervates these first four muscles. The trochlear nerve (CNIV) innervates the superior oblique and the abducens nerve (CNVI) innervates the lateral rectus.
 
 

Orbital Blood Vessels

The ophthalmic artery provides the chief blood supply for the orbital contents. It arises from the internal carotid artery as it leaves the cavernous sinus and passes through the optic canal within the dural sheath of the optic nerve. It then runs anteriorly, near the superomedial wall of the orbit and gives off branches for the structures in the orbit and the ethmoid bone.

The central retinal artery is a small, yet important branch of the ophthalmic artery, which arises inferior to the optic nerve and also runs inside the dural sheath of this nerve, until it reaches the eyeball. It then pierces the optic nerve and emerges at the optic disc and spreads over the internal surface of the retina to supply it.

The ciliary arteries are branches of the ophthalmic artery that supply the sclera, choroid, ciliary body and iris.

The lacrimal artery is another branch of the ophthalmic artery which supplies the lacrimal gland, conjunctiva and eyelids.

There are muscular branches of the ophthalmic artery which generally accompany branches of the oculomotor nerve.

There are branches of the ophthalmic artery which leave the orbit and anastomose with branches of the external carotid artery. These are: supraorbital; supratrochlear; dorsal nasal arteries. The anterior and posterior ethmoidal arteries also leave the orbit, however they enter the skull and terminate in the nasal mucosa.
 
 

The superior ophthalmic vein crosses superior to the optic nerve and passes through the superior orbital fissure to end in the cavernous sinus. It contain no valves, thus blood flow may be bidirectional, and anastomoses with the facial vein.

The inferior ophthalmic vein originates as a plexus in the floor of the orbit and passes through the inferior orbital fissure, after crossing inferior to the optic nerve, and ends in either the superior ophthalmic vein or the cavernous sinus directly.

The central retinal vein drains the internal structures of the eye and usually flows directly into the cavernous sinus.

The Visual Pathways

There is a spatial pattern of neural projection from the retina to the lateral geniculate nucleus, via the optic nerves and optic tracts, and from the lateral geniculate nucleus to the visual cortex of the occipital lobe, via the optic radiation.

Fibres from different parts of the retina represent different areas of the visual field, which is divided into nasal and temporal halves, and upper and lower quadrants, by vertical and horizontal lines through the fovea respectively. The macula lutea for central vision is represented separately in the visual pathway.

Fibres from the right half of the two retinae terminate in the right lateral geniculate nucleus and are relayed to the right visual cortex. The converse is true for the contralateral halves.

The fibres from the upper quadrants, excluding the macula lutea, project to the medial part of the lateral geniculate nucleus. These projections are then relayed to the anterior two-thirds of the visual cortex, above the calcarine sulcus. Conversely, the lower quadrant fibres are projected to the lateral part of the lateral geniculate nucleus and then to the anterior two-thirds of the visual cortex below the calcarine sulcus.

Projection fibres from the macula lutea run to a large posterior part of the lateral geniculate nucleus and are subsequently relayed to the posterior one-third of the visual cortex at the occipital pole.

The retinal image of an object in a visual field is represented in the cortex as inverted and reversed from right to left.

There are a small number of fibres that leave the optic nerve before reaching the lateral geniculate nucleus and terminate in the pre-tectal area and the superior colliculus. These fibres are involved in the pupillary light reflex. This reflex involves the constriction of the pupil in response to light being shone into the eye. Impulses of this light image are relayed to the superior colliculus and then to the Edinger-Westphal nucleus. This then relays impulses to the ciliary ganglion in the orbit and finally to the sphincter pupillae muscle in the iris. As a result of some fibres being sent across the midline in the posterior commissure to the contralateral Edinger-Westphal nucleus, the opposite pupil also constricts.

The accommodation-convergence reaction involves ocular convergence, pupillary constriction and thickening of the lens when focussing on a near object. This is clinically tested by asking the patient to look into the distance and then focus on an object about one foot away. When they are directing their attention to the near object, the medial recti muscles contract for convergence of the eyes, and the ciliary muscle causes thickening of the lens to increase its refractive power, and pupillary constriction to sharpen the image on the retina. For accommodation to occur, fibres from the visual association cortex travel to the midbrain via the superior brachium and terminate in the superior colliculus. From here the impulses are sent to the nuclei of the cranial nerves supplying the extra-ocular muscles and the Edinger-Westphal nucleus for the response of the
eyes. Thus the pathways for constriction of the pupils during this reflex and the pupillary light reflex are different.
.

 

HEARING

Detection of Sounds

To determine an observer's absolute threshold for sound presented through earphones, we measure the threshold sound pressure level called the minimum audible pressure. Minimum audible field refers to the intensity of a threshold stimulus measured in a free field rather than at the ear.

The ear is most senstive to sounds with freuencies between 1000 and 5000 Hz, being about 100 times ;less sensitive to a sound at 100Hz than to a sound at 3000 Hz. The most recent measurements given even lower minimum audible field for frequencies higher than 250 Hz, especially in ther region of 2000 to 4000 Hz.

Probably the most important reason for the lower thresholds in afree field is that the free-field situation allows resonance and amplification from the shape of the pinna and the ear canal.

The lower limit of sensitivity for the ear seems to be determined by the sound of blood rushing through the tiny vessels in the middle and inner ear, or perhaps be the random noise generated by motion of the air molecules where the upper limit is determined by the stimulus intensity that produces pain. The difference between the absolute threshold and the pain threshold for a particular frequency of sound waves defines the dynamic range of the ear for that frequency.

Threshold distance refers to the distance you can detect frequencies of about 16,000 Hz that decreases with age.

Temporal, Frequency, and Binaural Interactions
 

Several factors affect our ability to detect sounds:

1. frequency

2. intensity

3. duration

Temporal summation refers to a phenomenon by which the auditory system adds the sound energy received over about 20 msec for near-threshold stimuli giving rise to the relationship T = I X D. It states that any combination of intensity and duration that produces the same value will be heard with the same likelihood.

Just as there is a critical duration beyond which temporal summation did not occur, there is a critical band of frequencies beyond whihc adding tones does not facilitate detection. This critical band width is not the same width for all frequencies, being much narrower for low frequencies than for high frequencies.

The threshold for two-ear stimulation is about one-half of that for one-ear stimulation

When target and masker are presented at the same time, we have simultaneous masking. The greatest masking is found for tones that have frequencies similar to the masker itself and is greater for high frequency tones than for lower frequencies. The tones of low frequencies produce a very broad vibration pattern, extending over much of the membrane, whereas tones of higher frequencies produce vibration patterns nearer to the oval window and not extending as far along the membrane. The lower-freqeuncy target tone vibrates more of the membrane, the target's pattern extends beyond the flank of hte vibration pattern produced by the masker.

The upward spread of masking could be caused by the interaciton of the patterns of excitation produced on the basilar membrane by the target and masking sound. These vibration patterns would have their effects through stimulation of hte hair cells at the appropriate places on the basilar membrane. However, another mechanism could be contributing to the masking pattern, neurons responding to the masker could be suppressing the actual activity of neurons responding to the target tone.

If a masker is presented first, after some interstimulus interval by a brief target, any increase in the absolute threshold for the target is called forward masking. For interstimulus intervals longer than 300 msec, no forward masking occurs. The lower the frequency of both the target and mask pair, the more masking occurs. ALso, if the target tones are higher than the masker leads to greater masking.

Backward masking occurs maybe because the inhibition caused by the masker could build up faster than excitation caused by the target tone, thus overlapping with it in time and cancelling it to some extent, even when the target occurs appreciably earlier than the masker.

Central masking occurs whenr the target sound to one ear and a masking sound to another ear. When masker and target are presented to different ears, the masker must be about 50 dB more intense than when they are both presented to the same ear.

Informational masking occurs when a maksing sound is made up of several different frequencies chosen at random from trial to trial. When the mask is presented simultaneously with a target, the target is more difficult to detect even if none of the frequencies is particularly close to the target frequency. No informational masking occurs when the masking sound ends before the target tone begins. It has been estimated that about 20% of the simultaneous masking of tones is caused by informational masking. It might be caused by the inability to focus attention on the frequency of hte target in the presence of other sounds with frequencies in the same range.

Sound Discrimination

The problem of masking is really a problem of discrimination

The size of the Weber fraction for intensity differences is smallest for stimuli in the middle range of frequencies of about 1000 to 4000 Hz. Increasing or decreasing the frequency decreases our ability to discriminate intensity differences. The auditory system can detect changes of 10% to 20% in stimulus intensity. Intensity difference thresholds are about 33% smaller for binaural presentations. The threshold is also smaller the longer the duration of the stimuli over a range of 2 msec to 2 sec.

Profile analysis refers to the detection of an intensity icrement or other difference between complex sounds that differ in their profiles (amplitudes at the various frequencies making them up). Thresholds for signal frequencies between 500 and 2000 Hz can decrease by nearly 10 dB as the number of component frequencies is increased from 3 to 21. The auditory system seems to carry out a profile analysis in which the intensity of a tone at the signal frequency is compared with a weighted average of the intensities of the other components. Above 1000 Hz the Weber fraction is smallest about .005. binaural frequencies again are about 33% smaller.

Sound Localization

A number of cues indicate the direction:

1. the azimuth indicates the 0 deg straight ahead or 180 deg behind; the degree of left or right of the sound source. One ear receives the sound directly from the source while the other eare is in what could be called a sound shadow. Measurements have been made of the intensity differences between the ears as both the azimuth of the sound and the frequency of the emitted sound is varied.

2. time difference. for a sound at 90 deg azimuth in either direction, the ear closer to the sound is stimulated approx 0.80 msec earlier than the hidden ear.

3. phase difference referes to the difference in the phase of a sound wave between two ears caused by the different distances the sound wave has to travel to reach each ear; cue to localization of lower-frequency sounds. This is especially true for lower-frequency sounds because the time taken to compel one cycle is more than the maximum time difference of hte arrival of the sound at both ears. Although phase difference could be a cue to sound location, it provides ambigous information when we consider the full range of sound frequencies.

4. The pinnae delays or amplifies sounds at different frequencies by different amounts especially giving cues about elevation.

THE AUDITORY SYSTEM
 

SOUND

Sound is a form of mechanical pressure. Molecules in the air collide with others, causing air compression as the string moves forward and rarefaction as it moves back resulting in a wave of mechanical energy. Sound waves are alternations or rarefaction and compression of an elastic medium in which they travel and are created by rapid movements of a source in mechanical contact with this medium. The medium acquires some of the movement energy of the source and transfers it to other locations by means of collisions between the molecules of the medium. The wave tends to become less intense as it moves farther away from the original source. Sound cannot pass through a vacuum. The speed of sound varies according to the medium in which it travels. The speed of sound is fastest in water than in air.

The simplest sound wave is called sinusoidal

Wavelength is the distance from one peak of the wave to the next, representing a single cycle. The frequency of the wave, by convention, is the number of cycles the wave Is abfe tO ut half of their axons to the superior olive on the opposite side and half to the ipsilateral side. The cells of the dorsal cochlear nucleus send all their axons to the opposite side of the brain, eventually to terminate in the inferior colliculus. The two superior olives send afferent (sensory input) to the inferior colliculi. Special cells in the superior olive send efferent fibers to synapse with the hair cells in the cochlea. Most cells in the inferior colliculi send axons to the medial geniculate although a few go the superior colliculus perhaps to carry info about location. From the medial geniculate, fibers project to a part of the temporal cortex often called primary auditory projection area, or A1. An adjacent area, called A2 also receives axons. The vast majority of the auditory cortex is hidden in a fissure.

To understand the electrical activity, one must perform an auditory feature analysis
On neurons gave a burst of responses immediately after the onset of the tone bursts and then ceased responding, no matter how long the tone persisted.

Pauser neurons gave a similar burst of response, but this was followed by a pause and then a weaker, sustained response until the tone was turned off.

Choppers gave repeated bursts of firing followed by short pauses, with the vigor of successive bursts decreasing.

Primarylike neurons gave an initial vigorous burst of firing when the tone was turned on, and then the firing rate decayed to a lower level that was sustained for the duration of the tone.

Off neurons reduce their response rate below their spontaneous activity level at the onset of the tone and then give a burst of activity at its offset. For example, some cells have "W" shaped receptive fields of the medial geniculate.

 

Tonotopic arrangement of the basilar membrane up to the auditory cortex

Cells in the auditory cortex exhibit a variety of complex responses to sound stimuli. In 60% of the cells that respond to pure tones, there occur on responses, off responses, on-off responses, and more general excitatory and inhibitory responses. The other 40% of the cells seem to respond selectively to more complex sounds such as noise bursts. An interesting type of neuron in the auditory cortex is the frequency sweep detector that respond only to sounds that change frequency in a specific direction and range. Some cells respond only to increases in frequencies other only to decreases. A third type only response to increases in frequency for low-frequency tones.

Theory examined: "Triplex" extension of the "duplex" theory of pitch perception.

Triplex theory tries to account for three different ways that subjective pitch can arise from acoustic stimulation.

Theory assumes that the acoustic patterns delivered to the two ears are subjected to mechanical frequency analysis in the cochleas and that the products of that analysis are then subjected to a twofold coincidence or correlations analysis in unspecified centers of the nervous system (note: Heschl's gyrus in A1 of right temporal lobe).

Coincidence analyses:

1. Correlates the signals from the two ears and mediates sound localizations in phenomenal space.

2. Process of autocorrelation -- exposes periodicities that may have appeared on in the envelope, or the narrow-band signals of a form that is a smooth function of time that bounds the oscillations, and not necessarily in the waveform per se of the acoustic stimuli. The signals enter the two cochleas. The cochlear frequency analysis (F transformation) maps stimulus frequency into the spatial dimension, x, (see figure) and the ordinal relations along x are preserved in the excitation of the neurons of the auditory nerve (operation G). At each point along x, the wave is subjected to rectification, the effect of which is to "clip" off the negative swings of the oscillations. After it is rectified, the wave is then smoothed by the operation of a "low-pass filter." In the figure, the third wave is the wave that controls the excitation of the auditory neurons at x (i.e., the jth point on the x dimension of the basilar membrane).

While one neuron is "reloading", it neighbor may be firing. The overall effect down the auditory nerve is still a burst of activity for each pressure peak in the sound input, although different neurons are firing at different times. It is the composite response that makes up the frequency input at the higher neurological level.

The time-domain analyzer H preserves the order in x but adds an analysis of the y dimension, based mainly on interaural time differences, and an analysis in the z dimension, based on periodicities in the wave envelope received from FG. The neural signals are kept in their x channels, and corresponding right and left x channels pass each other as they cross the base of block H (i.e., place theory). The base of block H is shown schematically in figure below.

Neurons enter from left and right cochleas and make synaptic connections with the cell bodies (circles) of neurons ascending the z dimension. Temporally coinciding excitations are required to fire an ascending neuron. The y location of ascending activity therefore is dependent upon interaural time difference. Each small circle in figure represents the cell body of a neuron that projects upward and forms part of the "neuronal autocorrelator."

The straight through neuron is one of the ascending neurons in block H. With each straight through neuron is associated a number of delay chain neurons and a number of H-output (H-J) neurons. Several or many such networks operate simultaneously and in parallel on inputs that differ only in "microscopic" detail. From the assumption that, for an H-input neuron to fire, both fibers impinging upon it must fire within a short interval of temporal integration, it follows that the level of neural flux, regarded as a function jointly of time t and of the spatial coordinate z (which represents time delay T), is approximately the running autocorrelation function of the macroscopic (larger level of neuronal patterning) input.

In the projection from H to J, the x analysis is largely preserved, but each point in any frontal plane of H is, initially, connected to every point in the corresponding frontal plane of J. The H-J transformation organizes itself, according to rules imposed by the dynamics of the neuronal network, under the influence of acoustic stimulation. The patterns in J thereby acquire the properties that are reflected in pitch perception.

The basic condition for the emergence of clear subjective pitch is the activation of an isolated focus of tissue in J. The height II of the pitch depends upon the location of the focus. The definiteness of the pitch depends upon the degree of isolation, upon the remoteness of the active focus from other active tissue, upon the gradient of the distribution of activity.

Smell acts as if it has two separate modes of action that may result in different perceptual experiences and different forms of information extraction. One mode is associated with the experience of the flavors of food. The second is a distances sense, responding to molecules that float about in the air, carrying information about the objects or organisms from which they have become detached.

The stimuli must be volatile. Vibration frequency, the particular atoms that make up the molecule, and the number of electrons available for chemical bonding with other molecules on the smell receptors are also important.

The smell receptors cells that interact with he smell stimuli are located in the upper nasal passages called the olfactory epithelium or "smell skin". Each oval=shaped receptor cell sends a long extension, called an olfactory rod, to the surface of the olfactory epithelium, in addition sends axons toward the brain. The receptor cells function for about four to eight weeks before deteriorating. From a knob at the end of the olfactory rod protrude a number of olfactory cilia that are hairlike structures embedded in a special type of watery mucus secreted by a set of glands. The mucus contains many molecules of a special protein called olfactory binding protein (OBP), which can attach to hydrophobic odorant molecules that would ordinarily be repelled from the watery mucus. The olfactory cilia contain the receptor molecules that actually make contact with the smell stimulus. Humans have about 10 million olfactory receptors.

Probably at least two transduction mechanisms exist. One is made up of highly selective processes in which specific receptor cell proteins form reversible chemical bonds with specific parts of odorant molecules. These then cause depolarization of the receptor cells either directly or through other biochemical processes inside the cell. Numerous studies have suggested that a substance called cyclic adenosine monophospate (cAMP) is produced when an odorant molecule stimulates an olfactory receptor. A specific protein, called G_olf, that is activated by odorant reception and in turn activates a specific enzyme called adenylate cyclase which causes cAMP to be produced. The presence of cAMP causes the cell membrane to allow positively charged sodium ions to enter causing a depolarization. The second mechanism consists of less selective processes in which various chemicals directly affect the receptor cell membrane anywhere they contact it, causing ions to leak in or out, thus producing depolarization. These processes are the same for all receptor cells and may constituted what may be called generalist smell receptors. Humans likely possess both specialist and generalist receptors.

The lock and key theory maintains that variously shaped molecules fit into special sites on the receptor membrane like a key into a lock. The vibration theory maintains that the stimulus molecule ruptures certain chemical bonds in molecules making up a cell membrane, causing the release of stored-up energy, which in turn may trigger ion flow across the cell membrane either directly or indirectly. Which bonds are ruptured in which cells depends on the unique vibration frequency of each stimulus molecule.

The olfactory receptor cells send their axons through tiny holes in a bone at the top of the nasal cavity called the cribiform plate to form the olfactory nerve. The nerve goes straight to the olfactory bulb which is located in front of and below the main mass of the brain. The axons of the receptors and the dendrites of cells from the olfactory bulb form complex clusters of connections (called glomeruli) in the bulb grouped according to the type of receptor or type of stimulus molecule involved. One type of olfactory bulb cell sends axons to the primary sensory cortex for smell located in the temporal lobe. Another type sends axons to the limbic system as well as to the smell cortex. THe number of cells leaving the bulb are fewer than entering it. The major route of information from the olfactory bulb to the smell cortex is called the lateral olfactory tract with projections into the thalamus and other cortical areas.

The same sort of across-fiber patterns found in taste seem to be present in the olfactory system and it is possible the "code" for smell qualities can be found in these patterns. Some evidence has also indicated that odorant quality is coded as a pattern of activity across the entire olfactory bulb but a labeled-line theory such as the one for taste is probably too cumbersome.

When threshold stimuli were considered by Struiver (1960s), it was found that it takes 8 molecules at most to stimulate a single receptor cell in the human. Stimuli also varies across psychophysical methods, the purity of the odorant, delivery to the olfactory epithelium, and the intensity.

Anosmias refer to "odor-blindness" and there are 76 different anosmias.

Adaptation affects smell thresholds and the perceived intensity of an odor, and also the pleasantness of the odor. The largest loss of sensitivity was found for conditions of self-adaptation than for cross-adaptation. Cross-adaptation effects varied with the similarity of the smells of the two stimuli. Stimuli with similar smells gave large cross-adaptation effects.

Perceived odor intensity can be affected by the color of the stimulus or its water solubility. Stimulus discrimination of only 5% and smells can be smelled better through the right nostril that the left nostril and, since smell tracts stay on the ipsilateral side, may indicate right-hemisphere is more specialized for olfactory processing.

Henning's smell prism has not been confirmed by multidimensional scaling. The six "primary" odors defined by Henning are: Flowery, foul, fruity, burnt, spicy, and resinous. Amoore suggested there may be as many as 31 primary odors.

Alpha Androstenol, which is a sex-attractant pheromone for pigs and occurs in the human secretions. Most animals respond to two major types of pheremones: releasers, which on reception by an animal "release" a specific behavioral response and primers, which trigger glandular and other physiological activities in the recipient. For lower animals, pheremones elicit aggression in males, control estrus, the age at which puberty is attained, and even the likelihood of becoming impregnated. Copulins, female secretions of monkeys.

Social and behavior effects probably moderate the effects of pheremones in humans. People can detect their own body odors, and the odors of attractive v. unattractive persons, the sex of the odorant person. Females are better than males at discriminating.

Human sensitivity seems to present at an early age. Breast-feeding babies can recognize their own mother's breast and armpit odor.

What is needed is experience with the odor. Substance alpha androstenol has been demonstrated to increase the attractiveness of photographs and the prevalence of using a chair in a doctor's office. Whether human behavior can be controlled or not is debatable. However, women's menstrual cycles tend to be regularized and synchronized with another women in close proximity.

Taste Stimuli

The physical stimulus for the taste system are substances that can be dissolved in water. Four primary taste qualities: sweet, salty, sour, and bitter. A sweet taste is generally associated with so-called organic molecules, which are made up mostly of carbon, hydrogen, and oxygen in different combinations. Many sweet-tasting substances ahve a particular structure in common, termed the AB,H system, which consists of two negatively charged atoms. Bitter taste is related to sweet taste. Many substances that taste sweet in small amounts taste bitter in large amounts. Both also contain an AB,H system though in bitter substances the components have a different spatial arrangement. A salty taste is elicited by molecules that break into two electrical charged particles called ions. The size and weight of the negatively charged ions helps determine how salty a substance tastes. Substances with smaller negative ions taste saltier because the smaller negative ions can penertrate more easilty to the space around th etaste receptors, making it more electrically neutral and thus making it easier for the taste cell to signal the presence of the positive ion. Salts with large negative ions tend to taste bitter as well as sweet in large amounts but, in low doses, may taste sweet. Sour substances usually break up into acids. Hydrogen is the positively charged ion.

The tongue is covered with little bumps called papillae. The major receptors for taste are groups of cells called taste buds that are found in three types of papillae: Fungiform papillae, Foliate papillae, Circumvallate papillae, and Filiform papillae. Also, on the soft palate. Each taste bud consists of several receptor cells arranged like the closed petals of a flower. The mouth contains about 10,000 taste buds though the numbers decrease with age. Each taste bud has a span of a few days. The taste cells in the taste bud seem to be a specialized evolution of hte skin cells. A slender projection into th etop end of each cell lies near an opening onto the surface of the tongue called a taste pore. It is though that the actual reception mechanism for taste is located in these slender processes (called microvilli).

One theory states that salt ions, such as sodium, directly penetrate the cell mebrane of the microvilli, causing an electrical current to flow across the membrane. Another theory states that reversible chemical bonds form between specific parts of the taste stimulus molecule and receptor molecules in the cell membrane. These bonds change the flow of ions across the cell membrane, thus generating an electrical current. Finally, it is possible that some stimulus molecules alter the electrical properties of the cell membrane directly, causing an electrical current. For all of these, the final result is the release of a neurotransmitter across the synapse of the receptor cell with a taste nerve cell, causing spike potentials to travel up the taste nerves.

Three large nerves (the chorda tympani, vagus, and glossopharyngeal) carry fibers from the taste buds through the tongue to the solitary tract which is located in the medulla. Information is additionally carried from the common chemical sense which seperates from the taste system and in humans consists mostly of the trigeminal nerve of the head and its free nerve endings in the mouth and nasal cavity. From the nuclei in the solidary tract, taste information is carried via a set of pathways called the medial lemnisci to the taste center of the thalamus (ventral posterior nuclei). The thalamic taste area projects to three areas in the brain. Two are regions at the base of the primary somatosensory cortex and the third is the anterior-insular cortex which is a part of the frontal cortex under the front end of the temporal cortex.

Taste fibers respond to increasing intensity of the stimulus by increasing their overall rate of firing. A small increase in a weaker taste stimulus has a greater effect than a similar increase in a straonger taset stimulus. Most receptor cells seem to respond to all four primary tastes although with different senstivity. The code for taste quality is an across-fiber pattern of neural activity. THe labeled-line theory of taste quality encoding has each taste fiber encodes the intensity of a single basic taste quality, that associated with its best stimulus. This theory is compatible with the across-fiber pattern except that the code for taste quality is a profile across a few fiber types, rather than a pattern across many thousands of unique fibers. It is likely that some recoding takes place in the cortex. Specific cortical cells may given an "on" response to some taste stimuli and an "off" response to others.

Thresholds vary with viscosity or the temperature of the stimuli. Various parts of the mouth are not equally sensitive. Molar concentration is based on the weight of a substance dissolved in a given amount of a solvent. 1 mole is the molecular weight of the substance in grams added to enough water to make one liter of solution.

For the bitter substances, the lowest absolute threshold on the tongue is at the front and, even lower, on the palate. The tip and back of the tongue are most sensitive to sweet, while the front and sides are most sensitive to salt.

Taste blindness from PTC or vanillin or caffeine

Complete recovery can take up to two minutes

Adaptation to one substance can also have an effect on the threshold of another called cross-adaptation. In some cases, the exposure of the tongue to one stimulus may actually lower the threshold to another taste stimulus called potentiation.

"Off" responses to pure water has been found in the chorda tympani nerve because human saliva has salt and, therefore, already somewhat adapted to it.

The intensity of the taste sensation increases with the stimulus intensity on a linear relationship with magnitude estimations. It seems likely that our sensation of the intensity of a taste is directly related to the overall amount of neural activity evoked by the stimulus, which in turn depends on the intensity (Molar concentration) of the stimulus.

Henning depicted taste qualities in a three-dimensional pyramid with the primary tastes at the corners and has been confirmed by multidimensional scaling. RT also are similar for identification across similar tastes. Saltiness is fastest and bitterness is slowest, with sourness and sweetness intermediate, regardless of which salts, acids, etc. are utilized.

The Speech Stimulus

The ultimate goal of speech perception process it to put into consciousness an accurate representation of what the speaker intended to say. Depends on the analysis of speech in terms of how they are produced (phonetics) and how specific sounds distinguish words (phonemics).

Consonants and Vowels   Produces tow basic types of speech sounds: vowels and consonants. Closing mouth movements produce consonants whereas opening movements produce vowels. Consonants can be classified along three major dimension: 1. voiced or unvoiced 2. manner 3. place. Voiced consonants consist of a constriction of the flow of air out of the mouth followed very closely in time (less than 30msec) by vibration of the vocal folds (vocal chords). Unvoiced consonants, the vocal folds do not begin vibrating until a longer time after the constriction, usually more than 40 msec. The constriction of manner and place can be produced in three ways: stop consonants are formed by completely stopping the flow of air from the lungs and then suddenly releasing them "p" "t" in tea. Fricative are formed by stopping the flow through the nasal passages but leaving a small opening in the mouth and forcing air through it, producing some variety of "hissing" sound. Nasal consonants are produced through the nose. Most of the constrictions occur in two areas of the vocal tract. Labial constrictions are form from the lips of the lips against the teeth. The other area is inside the mouth. Here the tongue is positioned at various places, most of en at the ridge behind the teeth (alveolar), against the hard palate (palatal), or against the velum (velar). Articulators are the parts of the vocal tract used to produce speech sounds, such as teeth, etc. Vowels are produced differently from consonants. Depends: 1. position of tongue 2. degree of rounding of the lips whether rounded or unrounded.

Phonemes.  A basic unit of speech sound is the phone. A phone that is used to distinguish one word from another is called a phoneme. English has 40 basic phonemes It is difficult to find correspondence between acoustic features and perceived phonemes because of coarticulation or that we mover our articulators to produce sounds that provide information about several different phones simultaneously.

Speech spectogram .  The dark smudges in a speech spectrogram are called formants. Formants are bands of especially intense components that arise because the complex sounds waves created by the passage of air from the lungs across the vocal folds and out through the mouth or nose are affected by the positions of the various parts of the vocal tract. At least four formants are produced. The one at the lowest frequency (500 Hz)is called the first formant and is produced by the shape of the pharynx. The second formant, at about 1400 to 1500 Hz is produced by the shape of the oral cavity. Using a vocoder, only the first two formants are needed. Formants are centered on higher frequencies for the child and the adult male. Consonants are generally indicated by changes of formants over short intervals of time called formant transitions.

Our perception of the speech signal differs from its acoustical properties in that the acoustic signal is often a continuous stream without any obvious breaks or "markers" corresponding to those perceived subdivisions. Another feature the speech signal lacks is linearity. Linearity means that for each phoneme in an utterance, we should be about to find a corresponding segment and order of the physical speech signal. Another source of ambiguity comes from the signal lacking acoustic-phonetic invariance. Invariance refers to the notion that each phoneme must have some constant set of acoustic features associated with it whenever it is heard.

"Speech is special" versus "speech-is-just-another-form-of-auditory-pattern-perception"

At the physiological level, there is reason to suggest that speech is different from other aspects of auditory perception. Differences between the two hemispheres exists.

Categorical perception uses the continuum of voiced onset time for stop consonants. Phonemic boundary is around 35 msec. See page 428. Several studies have found similar results with nonspeech stimuli. Phonemic boundaries have been demonstrated in nonhuman species.

Category shift in that the effectiveness of the original stimuli has been weakened. Speech is special proponents stated that this was evidence that speech featrue detectors were present that were fatigued or adapted. Opponents obtained similar shifts in categorial boundaries with nonspeech stimuli, depending on the acoustic similarity of hte nonspeech stimulus used for adaptation to the speech stimulus used for testing.

Duplex perception in whihc the same sound can be perceived as having botht speech and nonspeech qualities. The most common technique used to deomonstrate this is to present a froup of synthetically generated formants, called the base stimulus, to one ear and present an isolated formant transition to the other ear. Typically, the base stimulus is heard as a particular syllable, say da, which either is heard this way or modified by the presence of hte formant transition in the other ear. Duplex perception happens when the sound is also heard as a "chirp" or othe rnonspeech sound. Can also be demonstrated for same ear presentations. The duplexity threshold, at which both sounds are heard, is about 20 dB higher than the threshold for discriminating speech sounds in the same stimuli. At least this indicates that speech processing dominates nonspeech processing though similar results have been found for musical chords and door slamming.

McGurk Effect where listening to a word and looking at a video produces a change in the perception of the word to a middle between the inconsistnet voiced words on the video and actual listened word. The McGurk effect is stronger for syllables than words.

Infants as young as 1 month of age discriminated speech stimuli better across phonemic boundaries than within phonemic categories, showing categorical perception in the way adults do. Babbling, especially in hearing imparied infants, suggests an early amodal language system. Young infants also deonstrate cross-modal integration of speech cues, even though they have had only limited experience with such cues. Infants also have a "language-general" discrimination ability, whereas adults ahve a "language-specific" discrimination ability.

homophones and homophenes (married and buried are different and sound different though produced by the same movements). Hear banket and lanket in each ear and hear blanket even though lanket proceeded banket by several milliseconds.

The more acoustic, syntactic, and semantic context is provided, the better the observers are at identifying the words. Phonemic restoration effect replaces a missing phoneme and is not perceived perceptually. The degree of phonemic restoration was related to whether or not visual cues were available as well as to the place and manner of articulation. It is also related to the number of possible completions for the word in the phrase that those that are unique.

The many theories of speech pereptioncan be subdivided on the basis of their level of analysis, whether they are oriented toward the identification of phonemens or words, and whether they utilze active or passive processing. Passive processing involves a filtering or feature detection sequence of events that is relatively fixed in nature. Active processing involves a much more interactiv sequence, in whihc the acoustic features are sensed, and then a series of higher-level processes involving analysis of the context (either meaning or phonetic) is considered.

Many passive theories incorporate the notions of feature detectors or template matching. Feature detectors for speech are usually conceptualized as neurons specialized for the detection of specific aspects of the speech signal. The concept of an auditory template may be viewed as a stored abstract represenation of certain aspects of speech that develops as a function of experience an dserves the same funciton as a feature detector. The use of feature detectors or templates in speech id is often views as a process similar to the visual feature extraction model called pandemonium, in that the signal is perceived on the basis of the template that it most closely matches or the phoneme or word that has the most features in common with it.

Some theories that are predominately passive in nature streess that ordinary auditory processes are sufficient to explain speech perception at the level of phonemes. These auditory theories usually postulate several stages of processing of speech sounds. The first stage consists of "ordinary" auditory processing, including analysis of complex sound into simple sine wave components, auditory feature analysis, and auditory pattern processing. For some theorists feature analysis comes first than patterna analysis, for others they occur at the same time in parallel. The next stage applies more specialized rules to the outputs of the first stages, integrating them to produce perception of phonemes. The Lexical Acess From Spectral (LAFS) model uses words rather than phonemes. In this model, the listener does a spectral analysis of the input signal, matching the results of this analysis to a set of templates of features stored in memory. Words are then identified from the set of features detected in the input. It is alos possible to have a passive model that use "special" speech units. The first stages would consist of detection of speech features by "special" feature detectors followed by integration of these features into precepts by "special" rules.

Active models are somewhat more variable, sicnce they often involve analysis of the context in which the speech is occuring, the expectations of the listener, the distribution of attentional resources, and memory components. Cohort theory involves early passive stages of analysis initially extract the first phoneme of a word. All words with phoneme are activated in memory and constituted the cohort. After the cohort is activated, other acoustic or phonetic information and higher-level expectancies operate to eliminate candidates. Phonetic refinement theory resembles cohort theory in its belief that initial feature processing activates candidate words that all sound alike along some dimensions and form a phonetic space. The set of candidates is narrowed until a best candidate is found. Words can be identified even if only partial information is available, sicnce even that partial information may activate the correct word more than it does any other word. Trace model by McClelland and Elman, begins with passive feature detection in three levels: (1) acoustic feature detectors whose output is the inputs to (2) phoneme detectors whose output is the input to (3) word detectors. The various detectors and other processors are nodes and are highly interconnected. Activating one node tends to activate all the nodes to which it is connected, both at the same level and at other higher or lower levels. All active theories have one thing in common: They all have higher-level decisional processes superimposed on the initial feature extraction results.

Motor Theory.  One of the most influential theories cannot be classified passive or active. Motor theory by Liberman et al. is oriented toward the id of phonemes considered as the intended phonetic gestures of a speaker; that is, what the speaker intended to say. The theory assumes that the same adaptations of the mammalian motor system that made speech possible for humans also made possible a system for perceiveing the sounds produced by the speech sounds. This system uses complex calculations to deduce the intended speech gestures fromt eh acoustic signal based on an abstract representation of the articulatory movements the listener would have used to produce just such a speech signal. A recent further development of this approach emphasizes the modular nature of the special and distinct system proposed to process speech sounds. A module consists of neural circuits that perform special processing to provide higher cognitive processes with representations of events that have particular biological, ecological, or behavioral significance.