Some experiments with computer synthesized speech

The generation of satisfactory artificial speech involves a great variety of semantic, linguistic, and articulatory factors. Many of these complex elements are as yet unknown, and computers are used extensively for their investigation. An account of an on-line, interactive computer facility used for such research at Bell Laboratories is given. The use of this facility for exploring the duration and pitch correlates of stress in English sentences is described.     

Statistical methods for a general theory of all-or-none learning

Greeno and Steiner have shown that a three state Markov Chain with a single absorbing state is equivalent to many of the current formalizations of All-or-None learning theories. Distribution statistics and other summary statistics are derived from the general model. Expressions for the maximum likelihood estimators of its parameters and the sampling variances of the estimates are presented. Likelihood ratio tests for several different null hypotheses are derived. These tests permit one to evaluate the usual null hypotheses in terms of the parameters of a process model.This research was supported in part by a USPHS Predoctoral Research Fellowship, 1-F1-MH-31-289-01, by a grant from the Graduate School of the University of Texas, and by grant HD-02212-01 from the National Institute of Child Health and Human Development to Dr. John Theios.     

A technique for the construction of discrete dichotic stimulation material

Procedures are described for the generation, synchronization. and accurate spacing of discrete dichotic stimulation material. A suggested standard definition of interpair interval is proposed.     

Detection and scaling of statistical differences between visual textures

Stochastically textured patterns were used to investigate the sensitivity of Ss to differences in the statistical distributions of locally defined properties of element density and shape. The results of the study indicate that when the textures were most structured, in terms of their variance, Ss were most accurate at detecting dissimilarities between pairs of patterns. Ss also rated the similarity of the statistical distributions of elements of patterns differing in local properties. Multidimensional scaling analysis of the ratings showed two dimensions, representing monotonic orderings of the stimuli.     

On the validity of scalp recordings of the auditory evoked potential

Simultaneous recording from scalp and cortical electrodes in unanesthetized and unrestrained cats revealed essentially no differences between the superficial and deep recordings of the auditory evoked responses. Human recordings taken under two conditions of muscular contraction contained no major differences from 30 to 300 msec after stimulus onset.     

Effects of deprivation and reinforcement-magnitude on response variability

Animals were trained to displace any one of five response keys in order to put themselves in a stimulus condition in which reinforcement could be obtained by depressing a response lever. Decreased deprivation and magnitude of reinforcement were found to increase the variability of the distribution of key responses. The relevance of these findings to other experiments in which deprivation, reinforcement magnitude, and intermittent reinforcement were studied is discussed.     

Several methods for teaching serial position sequences to monkeys

Three keys were available for monkeys to press. The objective was to teach the animals to press the keys in sequences up to 10 members in length. With fading procedures, a light that cued the correct key at a given serial member of the sequence faded out slightly each time the animal selected it, and became slightly brighter after the animal made an error at that sequence member. The correct keys were faded out, starting from the end of the sequence and proceeding toward the beginning. With control procedures, the cue lights were turned off suddenly, rather than being faded gradually. In almost every instance, the animals learned a longer series of unlighted key positions with the fading procedures than they did when each key-light was turned off suddenly. Also, requiring the animals to press a series of keys cued by lights before they could reach the sequence members they were to learn hampered them in learning the later serial members. By using several different sequences, it was possible to replicate these findings within the individual animal.     

The true distributions of the range of rank totals in the two-way classification

This paper presents the results from an investigation of the true probability distributions of the range of rank totals. A procedure for generating an approximation to the true distributions is also given. A comparison of the results of this approximation with an extensive criterion of generated true and sample distributions, and with other approximations is indicated. Accurate estimates of the critical ranges necessary to reach significance at three commonly used alpha levels, where the number of judges and items is less than or equal to sixteen, are presented in tabular form.