My ways to the studies of language were rather indirect. After having graduated in physics (1957) I read John von Neumann and Morgenstern’s Theory of Games and was fascinated by their exemplification of the modern axiomatic method. I asked myself the burning question how far mathematical theories and formalization could lead in disciplines that develop beyond the natural sciences. The question led me to studies of the humanities. I first concentrated on the philosophy of Leibniz’ de Arte Combinatoria and his Characteristica universalis and wrote my first dissertation thesis about Symbolic Representations used in modern science exemplifying, among other systems, networks of automata systems and the notations in Frege’s Conceptual Notation for Logic. Subsequently I studied Neo-Humboldtian linguistics and wrote my second thesis about its possible formalization in terms of information flow networks.

An interesting research position about theoretical and computational linguistics and their possible applications to machine translation led me to many cooperation visits to research institutes in Europe, the United States and Israel, and participation at the 1964 International Colloquium for Algebraic Linguistics and Automata Theory about linguistic models in Jerusalem. During my years in Berlin I formally compared the theoretical varieties of Generative Grammar with the more mathematical models of Montague Grammar. Changing from Berlin to the new University in Bochum initiated a new start, caused by organizing a colloquium in honour of the famous linguist R. Jakobson at the occasion of his honorary doctorate. Since Jakobson knew that our group had already studied the clear introduction and detailed descriptions to functional brain architecture in Popper-Eccles’ book The Self and its Brain he proposed the colloquium title: Language and Brain, hoping that we thus joined the new orientation he had described in a well-known New York University lecture.

Summary

Social grooming and rough-and-tumble play, along with caressing and hand-shaking, have something important in common, touching. Physical contact with another can be an essential ingredient of social communication – gentle touching can place the other animal at ease, whereas rough contact can do the opposite. Although the underlying neurobiology is still to be fully mapped, it does appear that there is a common set of neurochemical pathways that regulate these touch-induced changes in mood across mammals. Given its potential value in the manipulation of the affective state of social partners, it should not be surprising that touch is an important component of communication. A close analysis of the comparative and neurobiological literature on rough-and-tumble play, or play fighting, suggests that there are two levels of control over this touch-based communication. Firstly, there is the subcortically regulated emotional state of the interactants. Secondly, there is the cortically mediated modulation of the touching behavior that allows animals to use touch in a more strategic manner. How these two levels interact and what social conditions foster the need for additional cortical control over touch remains to be determined.

Introduction

A hostile donkey is rendered peaceful by the human object of its ire vigorously rubbing the base of its tail (Ewer 1967), an anxious monkey is calmed down after being groomed by a social partner (Goosen 1981), and agitated rats relax after social play (Arelis 2006; Darwish et al. 2001). What do all these situations have in common?

Introduction

The simplest way to study learning is to expose subjects to a stimulus and then assess whether they show some effect which is absent in subjects lacking that experience (see Rescorla, 1998). One stimulus exposure effect is latent inhibition. Subjects in one group but not another are exposed to a stimulus in the absence of any other scheduled event. Then subjects in both groups are exposed to a signaling relation between that stimulus (the conditioned stimulus (CS)) and a motivationally significant event (an unconditioned stimulus (US)). The responding elicited by the CS in subjects just exposed to the signaling relation is depressed in those pre-exposed to the stimulus. Conditioned responding is said to have been latently inhibited by the prior stimulus-alone exposures. This effect has also been observed in a within-subject design where subjects are first exposed to one stimulus but not to another and then to a signaling relation between each of these stimuli and a US. Responding develops more rapidly to the novel CS than to one that had been pre-exposed (e.g., Killcross & Robbins, 1993; Rescorla, 2002a, 2002b).

Another effect of stimulus exposures is extinction. Two groups of subjects are exposed to a signaling relation between a CS and US. Then subjects in one group but not the other are exposed to the CS in the absence of any other scheduled event. The responding elicited by the CS in subjects just exposed to the signaling relation is depressed in those that additionally received the CS-alone exposures.

Latent inhibition (LI) is a robust phenomenon in which repeated preexposure to a stimulus that is not reinforced retards future associability to that stimulus (Lubow,1989). LI has been uniformly accepted as an adaptive mechanism across a variety of species (Lubow & Gewirtz, 1995). In humans, a deficit in LI has been associated with the active phase of schizophrenia and with psychosis-proneness (Baruch, Hemsley, & Gray, 1988a, 1988b; Lubow, Ingberg-Sachs, Zalstein-Orda, & Gewirtz, 1992). However, attenuated LI has also been reported in non-disordered normal subjects (see Braunstein-Bercovitz, Rammsayer, Gibbons, & Lubow, 2002), suggesting that attenuated LI exists on a continuum that extends from hospitalized psychotics to high-functioning normals. Recent research suggests that there may be situations in which attenuated LI actually confers an advantage to individuals. There is a growing body of evidence that indicates attenuated LI may be present in a subset of high-functioning and creative individuals (e.g. Carson, Peterson, & Higgins, 2003; Peterson & Carson, 2000). Attenuated LI may increase the probability of making novel or original associations among disparate stimuli by increasing the amount of information available to conscious awareness.

The theoretical relationship of creativity and latent inhibition

Latent inhibition is widely accepted as an index of the ability to ignore irrelevant stimuli (Lubow & Kaplan, 2005). When latent inhibition is expressed, the ability to form associations to information deemed irrelevant is reduced. Conversely, when latent inhibition is attenuated, the ability to form associations to seemingly irrelevant information is expanded.

What one knows and what one shows: acquisition vs. performance

Since the time of Aristotle, it has been commonly assumed that two events occurring in close proximity become associated to each other. This learning by contiguity is a central determinant of the phenomenon of classical conditioning, in which a neutral target stimulus (X) is repeatedly presented in close proximity to a stimulus that unconditionally produces a response (i.e., an unconditioned stimulus, US), with the consequence that X comes to elicit a response appropriate to the US. That is, X becomes a conditioned stimulus (CS) that produces a conditioned response (CR). However, there have been reports of various conditions under which CS–US pairings fail to result in an acquired response to the CS. For example, even though pairings of X and the US ordinarily result in a robust CR (i.e., acquisition), X will fail to gain response potential if it is trained in the presence of another, more salient CS, A (i.e., overshadowing; Pavlov,1927). Thus, even though the contiguity between CS X and the US is the same in the X–US acquisition and AX–US overshadowing conditions, the resulting behavioral control by CS X is not. An easy explanation for this difference is that subjects did not acquire an X–US association in the overshadowing condition.

The ability to adapt to environmental change is essential for the survival of any organism. Adaptations to environmental change include the ability to learn associations and the ability to modify those associations. Numerous studies have demonstrated that deficits in the ability to modify or modulate learning are associated with multiple mental illnesses and disorders (Amieva, Phillips, Della, & Henry, 2004; Baron-Cohen & Belmonte, 2005; Baruch, Hemsley, & Gray, 1988; Clark & Goodwin, 2004; Kaplan et al., 2006; Lubow & Gewirtz, 1995; Lubow & Josman, 1993; Vaitl, Lipp, Bauer et al., 1999; Weiner, Schiller, & Gaisler-Salomon, 2003). Because of the strong link between mental illness and deficits in processes that modulate learning, understanding the neural substrates of these processes could facilitate development of treatments for many diseases. Latent inhibition is one process that can modulate learned associations and is also altered in patients with mental illness (see the chapters on schizophrenia in this book for an in-depth discussion).

As described in preceding chapters in this book, latent inhibition is the phenomenon in which pre-exposure to a conditioned stimulus (CS) prior to the pairing of this CS with an unconditioned stimulus (US) decreases the subsequent conditioned responses (CR). There are three phases of latent inhibition: (1) pre-exposure to the CS, (2) training (or conditioning), and (3) testing. The presence of latent inhibition is identified by comparing the degree of conditioned responding between the CS pre-exposed group and the non-pre-exposed group.

Summary

It is clear that the concept of latent inhibition (LI) and the notion that it might be abnormal in schizophrenia (SZ) patients have been powerful heuristic tools for cross-species studies. Less clear has been the evidence that LI is actually abnormal in SZ, and if so, what the nature of such an abnormality might be, and which types of SZ patients might manifest it. We previously reported in two studies our ability to detect normal LI in a total of 88 SZ patients who successfully learned the non-preexposure task. Normal LI in these subjects could not be easily explained by peculiarities of the design of the LI task, or the characteristics of the study sample. Since submission of the last of these reports in 2004, we identified a total of three Medline papers in which LI was tested in SZ patients: one reported reduced LI only in unmedicated patients with predominant positive symptoms, another found elevated LI in only 6 out of 30 predominantly medicated patients who had the combination of low positive symptoms and high negative symptoms, and the third reported that LI was both reduced and elevated at different times within a single test, among medicated patients, unrelated to positive or negative symptoms. The belief that LI is abnormal in SZ persists, despite a paucity of clear, replicated, direct supportive evidence, and despite the presence of substantial relevant information that might lead us to conclude otherwise.

Introduction

Inspired by early Darwinian theory, cross-species comparisons of learning abilities were once a focal point of experimental psychology. Today, many researchers have turned to simple organic systems, not to compare them to other species, but rather to take advantage of the relative lack of complexity of their cellular and molecular architecture for the purpose of modeling basic processes that are assumed to be operative in more intricate organisms. This approach has been used with a number of invertebrates, including the fruit fly Drosophila, the sea slug Aplysyia, and the honey bee Apis mellifera (for reviews, see, e.g., Davis, 2005; Kandel, 2001; Menzel & Muller, 1996, respectively).

Many of these studies have examined the neural pathways involved in classical conditioning. More recently, however, attention also has been directed to an even simpler behavioral phenomenon, at least operationally, namely latent inhibition (LI). In the classical conditioning paradigm, the subject encounters paired presentations of the CS and US and the experimenter records changes in responsivity to the CS (i.e., CRs). In LI, the subject is first presented with a series of to-be-CSs, each of which is not followed by an event of consequence (CS−0). Typically, responses to the to-be-CS are not documented. The stimulus preexposure stage is followed by one or more CS−US pairings, during which time CRs are recorded (two-stage procedure). Alternatively, the second stage may be followed by an additional stage in which the CSs again are presented without the US, and the CRs are monitored (three-stage procedure).

Latent inhibition refers to the observation that under specific conditions, nonreinforced preexposure (PE) to a stimulus retards the efficacy with which this stimulus is conditioned when paired with reinforcement, compared to a nonpreexposed (NPE) stimulus. The pharmacology of LI has been almost exclusively associated with the use of LI as an animal model of schizophrenia, and therefore largely overlaps the pharmacology of schizophrenia.

As detailed in our chapter on neural substrates of LI, the widely held notion that nonreinforced stimulus preexposure reduces attention to or salience of stimuli has served to link LI to attentional processing in schizophrenia. Specifically, because schizophrenia is characterized by an inability to filter out, or ignore, irrelevant or unimportant stimuli, abnormal LI was proposed as a tool for modeling deficient attention in schizophrenia (Solomon, Crider, Winkelman et al.,1981; Weiner, Lubow, & Feldon, 1981, 1984, 1988). Here we will focus on our use of LI for the development of pharmacological animal models related to schizophrenia and the identification of viable antipsychotic/anti-schizophrenia medications. Based on our initial pharmacological data, we adopted in our pharmacological investigations a view of LI that distinguished between the acquisition of LI (learning to ignore the nonreinforced stimulus in preexposure) and the expression of LI (subsequent expression of this learning in conditioning) (Weiner, Feldon, & Katz, 1987; Weiner et al., 1984, 1988). This view of LI has been elaborated in the switching model of LI (Weiner, 1990, 2003; Weiner & Feldon, 1997), and has guided our use of LI for modeling schizophrenia.

Introduction: effects of time and context on latent inhibition

Despite its apparent simplicity, the phenomenon of latent inhibition (LI) represents one of the most sophisticated and flexible mechanisms that organisms with complex nervous systems have developed through evolution to ensure efficient interaction with the environment. Because the environment is constantly changing, mechanisms that determine the processing of a neutral stimulus depend on a large range of different circumstances. In this chapter we will focus on the role played by two factors, namely, time and context, that seemingly affect LI separately, as well as in combination. The impact of these factors (both apart and conjointly) on LI is still one of the greatest challenges to associative theories of learning.

In any learning process there is a series of elements that determine the intensity and type of association that is formed. In the case of classical conditioning, some of the parameters on which Pavlov (1927) concentrated in his original studies were related to the temporal contiguity between stimuli (e.g., whether the stimuli involved in the pairings were presented simultaneously or sequentially, or the order in their presentation when presented sequentially), as well as to the excitatory vs. inhibitory nature of the association acquired under different treatments. Some other elements that have subsequently demonstrated their relevance to associative learning were also pointed out by Pavlov, although sometimes in a quite intuitive manner. For example, he mentioned that conditioned reflexes could be affected by the surrounding stimuli during conditioning in the animal's environment.