Two ways of learning associations

How people learn chunks or associations between adjacent items in sequences was modelled. Two previously successful models of how people learn artificial grammars were contrasted: the CCN, a network version of the competitive chunker of Servan‐Schreiber and Anderson [J. Exp. Psychol.: Learn. Mem. Cogn. 16 (1990) 592], which produces local and compositionally‐structured chunk representations acquired incrementally; and the simple recurrent network (SRN) of Elman [Cogn. Sci. 14 (1990) 179], which acquires distributed representations through error correction. The models' susceptibility to two types of interference was determined: prediction conflicts, in which a given letter can predict two other letters that appear next with an unequal frequency; and retroactive interference, in which the prediction made by a letter changes in the second half of training. The predictions of the models were determined by exploring parameter space and seeing howdensely different regions of the space of possible experimental outcomes were populated by model outcomes. For both types of interference, human data fell squarely in regions characteristic of CCN performance but not characteristic of SRN performance.     

THE GOLEM LEGEND AND THE ENIGMA OF FACEBOOK

We are easily misguided as to the true nature of Facebook, and tend to treat it simply as a powerful technological instrument in the service of human intentions. We can, however, gain a better picture of it through recourse to the Jewish tradition of the golem, an image of human beings, created by them in a re-enactment of their own creation by God. It turns into a magic servant in modernity with an inherent dynamic running between its human and its subhuman characteristics. This dynamic is the main cause behind its becoming uncontrollable. In like manner, what is subhuman in Facebook serves its masters and functions under their total control, but also empowers Facebook's increasingly human operation, an algorithm-based capability which raises growing doubts about what counts as human. Facebook implies the crisis of humanity which coincides with the “death of God,” that is, the obsolescence of the idea of a divine creator.     

Does nondirectional signalization of target distance contribute to navigation in the Morris water maze?

The possibility that rats can navigate in the Morris water maze by reducing the difference between the memorized platform scene and the current sensory input was tested in nine blind rats. A computerized videosystem monitored the rats' movements in the pool and converted the rat-target distance into tones the frequency of which increased in 64 equal steps from 120 Hz at 128 cm to 7680 Hz at 0 cm. During 15 days of training to find a fixed platform position from different starting points (12 trials per day) average escape latencies decreased from 39.0 to 25.4 s. The performance significantly deteriorated when the acoustic distance signalization was omitted and/or when the target position was changed form trial to trial. It is concluded that blind rats solved the task by simultaneously employing search strategy based on position responses, mapping using acoustic background beacons, and distance reduction navigation. It is argued that the various strategies are additive and that their relative significance depends of the conditions of the experiment.     

Brain weights correlate with behavioral parameters in individual inbred mice housed in a common and enriched environment

Fifty 8-week-old Balb/c mice were individually identified and housed together in a large and enriched environment for 5 months. Maze and open field exploration, response to an aversive noise, swimming, and induced grooming tests were applied to each mouse in an initial search for possible relationships between brain morphology and spontaneous behavior in isogenic individuals living in a complex social and physical environment. The tasks generated 39 quantitative behavioral indices which include locomotion, rearing, still, and grooming bout frequencies, latencies, total, and mean bout durations. At the end of the tests, the 7-month-old mice were sacrificed and the fresh weights of their whole brain, cerebellum, brain stem, diencephalon, telencephalon, and prosencephalon were rapidly obtained. Behavioral data have wide variations and do not adjust to normal population distributions. Means of the same parameter differ between tests. A Spearman correlation matrix of all data yielded many significant correlations between indices of the same task which can be interpreted in terms of time budget and sequence probability. Significant correlations between indices of different tests suggest diverse emotionalities, exploratory strategies, and motor skills. The correlations between body and brain weights and among separate brain regions were not significant. There were several low but significant correlations between brain weights and behavioral indices. Such correlations, the resulting factors, and significant behavioral differences between mice with large and small brains suggest that mice displaying low motor activity in novel environments have larger brains and forebrain/hindbrain ratios than mice with high activity, and that animals with high scores of some specific behaviors have larger brain areas physiologically related to such behaviors.(ABSTRACT TRUNCATED AT 250 WORDS)