Parallel Distributed Processing and Semantic Cognition

Timothy T. Rogers and James L. McClelland have distilled the essence of the arguments in their book "Semantic Cognition: A Parallel Distributed Processing Approach" for an article (PDF here) to appear in Brain and Behavioral Sciences with peer commentary. Here is their abstract:

In our recent book, we present a parallel distributed processing theory of the acquisition, representation and use of human semantic knowledge. The theory proposes that semantic abilities arise from the flow of activation amongst simple, neuron-like processing units, as governed by the strengths of interconnecting weights; and that acquisition of new semantic information involves the gradual adjustment of weights in the system in response to experience. These simple ideas explain a wide range of empirical phenomena from studies of categorization, lexical acquisition, and disordered semantic cognition. In this précis we focus on phenomena central to the reaction against similarity-based theories that arose in the 1980's and that subsequently motivated the "theory-theory" approach to semantic knowledge. Specifically, we consider i) how concepts differentiate in early development, ii) why some groupings of items seem to form "good" or coherent categories while others do not, iii) why different properties seem central or important to different concepts, iv) why children and adults sometimes attest to beliefs that seem to contradict their direct experience, v) how concepts reorganize between the ages of 4 and 10, and vi) the relationship between causal knowledge and semantic knowledge. The explanations for these phenomena are illustrated with reference to a simple feed-forward connectionist model; and the relationship between this simple model, the broader theory, and more general issues in cognitive science are discussed.

In an Ultimatum Game, Chimps, but not humans, are rational maximizers

Another interesting bit of work from Tomasello's group (PDF here), in which they devised an ingenious apparatus for a mini-ultimatum game, a reduced form of the ultimatum game in which proposers are given a choice between making one of two pre-set offers which the responder can then accept or reject. The proposer had as one option an amount that would typically be rejected by a human responder as unfair, namely 80% for the proposer and 20% for the responder. The most important finding was that responders tended to accept any offer. These results support the hypothesis that other-regarding preferences and aversion to inequitable outcomes, which play key roles in human social organization, distinguish us from our closest living relatives. Here is their abstract, slightly edited:

Traditional models of economic decision-making assume that people are self-interested rational maximizers. Empirical research has demonstrated, however, that people will take into account the interests of others and are sensitive to norms of cooperation and fairness. In one of the most robust tests of this finding, the ultimatum game, individuals will reject a proposed division of a monetary windfall, at a cost to themselves, if they perceive it as unfair. Here we show that in an ultimatum game, humans' closest living relatives, chimpanzees (Pan troglodytes), are rational maximizers and are not sensitive to fairness.

Figure: Illustration of the testing environment. The proposer, who makes the first choice, sits to the responder's left. The apparatus, which has two sliding trays connected by a single rope, is outside of the cages. (A) By first sliding a Plexiglas panel (not shown) to access one rope end and by then pulling it, the proposer draws one of the baited trays halfway toward the two subjects. (B) The responder can then pull the attached rod, now within reach, to bring the proposed food tray to the cage mesh so that (C) both subjects can eat from their respective food dishes (clearly separated by a translucent divider).

Interactions between Declarative and Procedural Memories

Brown and Robertson do a simple experiment showing interaction between two main memory systems usually thought to be independent. Their abstract, slightly edited:

The acquisition of declarative (i.e., facts) and procedural (i.e., skills) memories may be supported by independent systems. This same organization may exist, after memory acquisition, when memories are processed off-line during consolidation. Alternatively, memory consolidation may be supported by interactive systems. This latter interactive organization predicts interference between declarative and procedural memories. Here, we show that procedural consolidation, expressed as an off-line motor skill improvement, can be blocked by declarative learning over wake, but not over a night of sleep. [note: the procedural task was learning a sequence of visually cued button presses at four possible postions.] The extent of the blockade on procedural consolidation was correlated to participants' declarative word recall. Similarly, in another experiment, the reciprocal relationship was found: declarative consolidation was blocked by procedural learning over wake, but not over a night of sleep. [note: The declarative task was learning a list of 16 words each individually presented on a computer screen for two seconds, with the list being presented in the same order five times.] The decrease in declarative recall was correlated to participants' procedural learning. These results challenge the concept of fixed independent memory systems; instead, they suggest a dynamic relationship, modulated by when consolidation takes place, allowing at times for a reciprocal interaction between memory systems.

Facts are no match for gossip...

In yesterday's Science NY Times, John Tierny describes work by Ralf D. Sommerfeld and colleagues on human reciprocity that has a surprising result (PDF here). Here are some clips:

Language, according to the anthropologist Robin Dunbar, evolved because gossip is a more efficient version of the “social grooming” essential for animals to live in groups...Gossip also told people whom to trust, and the prospect of a bad reputation discouraged them from acting selfishly, so large groups could peacefully cooperate. At least, that was the theory: gossip promoted the “indirect reciprocity” that made human society possible.

To test it, researchers at the Max Planck Institute for Evolutionary Biology and the University of Vienna gave 10 Euros apiece to 126 students and had them play a game that put them in a dilemma. On each turn, the players would be paired off, and one of them was offered a chance to give 1.25 Euros to the other. If he agreed, the researchers added a bonus of .75 Euro so that the recipient ended up gaining 2 Euros...If the first player refused to give the money, he’d save 1.25 Euros, but if others found out about his miserliness they might later withhold money from him. As the game progressed, with the players changing partners frequently and alternating between the donor and recipient roles, the players were given information about their partners’ past decisions...Sometimes the donor was shown a record of what the partner had done previously while playing the donor role. The more generous this partner had earlier been toward other players, the more likely the donor was to give him something...Sometimes the donor was shown gossip about the partner from another player. When the partner was paid a compliment like “spendabler spieler!” — generous player! — the donor was more likely to give money. But the donor turned stingy when he saw gossip like “übler geizkragen” — nasty miser.

So far, so good. As predicted, gossip promoted indirect reciprocity. The research, published in the Proceedings of the National Academy of Sciences [Note, I haven't found this article yet in a PNAS search], showed that most people passed on accurate gossip and used it for the common good. They rewarded cooperative behavior even when they themselves weren’t directly affected by the behavior...If a cooperation game like this was played without consequences for the players’ reputations — as has been done in other experiments — most players would be miserly, and cooperation would collapse. In this experiment they were generous most of the time, and on average ended up with twice as much money as they had at the beginning of the game.

But here’s the disconcerting news from the experiment. In a couple of rounds, each donor was given both hard facts and gossip. He was given a record of how his partner had behaved previously as well as some gossip — positive gossip in one round, negative in another...The donor was told that the source of the gossip didn’t have any extra information beyond what the donor could already see for himself. Yet the gossip, whether positive or negative, still had a big influence on the donors’ decisions, and it didn’t even matter if the source of the gossip had a good reputation himself. On average, cooperation increased by about 20 percent if the gossip was good, and fell by 20 percent if the gossip was negative.

Why is this?

It could be,” Sommerfeld suggested, “that we are just more adapted to listen to other information than to observe people, because most of the time we’re not able to observe how other people are behaving. Thus we might believe we have missed something."

Most popular consciousness papers for September 2007

From the ASSC Eprints Archive:

1. Rosenthal, David (2007) Consciousness and its function. In: 11th annual
meeting of the Association for the Scientific Study of Consciousness, 22-25
June 2007, Las Vegas, USA. 1014 downloads from 20 countries.
http://eprints.assc.caltech.edu/293/
2. Koriat, A. (2006) Metacognition and Consciousness. In: Cambridge handbook
of consciousness. Cambridge University Press, New York, USA. 870 downloads
from 18 countries. http://eprints.assc.caltech.edu/175/
3. Windt, Jennifer Michelle and Metzinger, Thomas (2006) The philosophy of
dreaming and self-consciousness: What happens to the experiential subject
during the dream state? In: The new science of dreaming. 834 downloads from
16 countries. http://eprints.assc.caltech.edu/200/
4. Sagiv, Noam and Ward, Jamie (2006) Crossmodal interactions: lessons from
synesthesia. In: Visual Perception, Part 2. Progress in Brain Research,
Volume 155. 814 downloads from 14 countries.
http://eprints.assc.caltech.edu/224/
5. Droege, Paula (2007) ASSC-11 Conference Report. In: 11th annual meeting
of the Association for the Scientific Study of Consciousness, June 22-25,
2007, Las Vegas, Nevada. 798 downloads from 18 countries.
http://eprints.assc.caltech.edu/323/

Motivation alters physical perception

Here is a clip from the Editor's choice section of the Oct. 5 issue of Science:

Students in elementary physics classes are introduced to the concept of frame of reference--the spatial coordinate system used by an observer to describe events--for instance, in the context of the perceived motion of trees by a passenger in a moving automobile. Adding in the dimension of time leads into non-intuitive territory, as in the example of a traveling astronaut who returns to Earth younger than her stay-at-home twin.

Building on previous work that demonstrated that internal physiological states can influence one's perception of physical quantities (such as thirsty people being more likely to characterize objects as transparent; that is, resembling water), Balcetis and Dunning show that internal psychological states are also capable of altering our perception of the external world. They induced states of high or low cognitive dissonance (a mismatch between thought and action) by asking or telling two groups of students to walk across campus wearing various fruit- and vegetable-themed adornments. In order to render a freely chosen yet somewhat embarrassing task less unpleasant to fulfill, the first set of students mentally shortened the distance they had to cover by estimating it to be fully 40% less than the average estimate made by the second group. Intriguingly, the route to ameliorating the state of dissonance appeared to be purely perceptual, as the free-choice students did not shorten their time of exposure by walking faster; in fact, they took about 10% longer.

The abstract of the original article;

Two studies demonstrated that the motivation to resolve cognitive dissonance affects the visual perception of physical environments. In Study 1, subjects crossed a campus quadrangle wearing a costume reminiscent of Carmen Miranda. In Study 2, subjects pushed themselves up a hill while kneeling on a skateboard. Subjects performed either task under a high-choice, low-choice, or control condition. Subjects in the high-choice conditions, presumably to resolve dissonance, perceived the environment to be less aversive than did subjects in the low-choice and control conditions, seeing a shorter distance to travel (Study 1) and a shallower slope to climb (Study 2). These studies suggest that the impact of motivational states extends from social judgment down into perceptual processes.

Baboon Metaphysics

Nicholas Wade comments (PDF here) on the work of Cheney and Seyfarth, whose most recent book has the title of this post. They have carried out an ingenious series of experiments in the wild to probe how the baboon's mind keeps track of transient relationships:

Royal is a cantankerous old male baboon whose troop of some 80 members lives in the Moremi Game Reserve in Botswana. A perplexing event is about to disturb his day....From the bushes to his right, he hears a staccato whoop, the distinctive call that female baboons always make after mating. He recognizes the voice as that of Jackalberry, the current consort of Cassius, a male who outranks Royal in the strict hierarchy of male baboons. No hope of sex today....But then, surprisingly, he hears Cassius’s signature greeting grunt to his left. His puzzlement is plain on the video made of his reaction. You can almost see the wheels turn slowly in his head:...“Jackalberry here, but Cassius over there. Hmm, Jackalberry must be hooking up with some one else. But that means Cassius has left her unguarded. Say what — this is my big chance!”...The video shows him loping off in the direction of Jackalberry’s whoop. But all that he will find is the loudspeaker from which researchers have played Jackalberry’s recorded call.

Although Baboons excel at the skills required for maintaining social networks regulated by matrilineal lines and dominance hierarchies, there is no evidence that they attribute beliefs or ideas to other animals, or that 'they know that they know.' They provide an example of what sort of social and cognitive complexity is possible in the absence of language and a theory of mind.

The prospective brain

I've been meaning to point you to a nice review article by Daniel Schacter and colleagues (PDF here). Here is the abstract and a summary figure.

A rapidly growing number of recent studies show that imagining the future depends on much of the same neural machinery that is needed for remembering the past. These findings have led to the concept of the prospective brain; an idea that a crucial function of the brain is to use stored information to imagine, simulate and predict possible future events. We suggest that processes such as memory can be productively re-conceptualized in light of this idea.

The core brain system that is consistently activated while remembering the past, envisioning the future and during related forms of mental simulation is illustrated schematically. Prominent components of this network include medial prefrontal regions, posterior regions in the medial and lateral parietal cortex (extending into the precuneus and the retrosplenial cortex), the lateral temporal cortex and the medial temporal lobe. Moreover, regions within this core brain system are functionally correlated with each other and, prominently, with the hippocampal formation. We suggest that this core brain system functions adaptively to integrate information about relationships and associations from past experiences, in order to construct mental simulations about possible future events.

New phases of our lives...

NY Times columnist David Brooks muses on how the phases of our life have increased in number (PDF here). Some clips:

There used to be four common life phases: childhood, adolescence, adulthood and old age. Now, there are at least six: childhood, adolescence, odyssey, adulthood, active retirement and old age. Of the new ones, the least understood is odyssey, the decade of wandering that frequently occurs between adolescence and adulthood...People who were born before 1964 tend to define adulthood by certain accomplishments — moving away from home, becoming financially independent, getting married and starting a family....In 1960, roughly 70 percent of 30-year-olds had achieved these things. By 2000, fewer than 40 percent of 30-year-olds had done the same.

...you can see the spirit of fluidity that now characterizes this stage....everything seems to give way to a less permanent version of itself...Dating gives way to Facebook and hooking up. Marriage gives way to cohabitation....Graduating seniors don’t find corporations offering them jobs that will guide them all the way to retirement. Instead they find a vast menu of information economy options...what we’re seeing is the creation of a new life phase, just as adolescence came into being a century ago...European nations are traveling this route ahead of us... Europeans delay marriage even longer than we do and spend even more years shifting between the job market and higher education...Someday people will look back and wonder at the vast social changes wrought by the emerging social group that saw their situations first captured by “Friends” and later by “Knocked Up.”

How to get health facts completely wrong....

John Tierney reviews Gary Taubes account (in "“Good Calories, Bad Calories”; Knopf, 2007) of how the health science establishment got it completely wrong, from the basic science on through official government pronouncements on dietary fat and hearth disease (PDF here). Some clips:

The notion that fatty foods shorten your life began as a hypothesis based on dubious assumptions and data; when scientists tried to confirm it they failed repeatedly.

Even so, in 1988, the surgeon general, C. Everett Koop, proclaimed ice cream to a be public-health menace right up there with cigarettes.This happened because of what social scientists call a cascade:

We like to think that people improve their judgment by putting their minds together, and sometimes they do. The studio audience at “Who Wants to Be a Millionaire” usually votes for the right answer. But suppose, instead of the audience members voting silently in unison, they voted out loud one after another. And suppose the first person gets it wrong...If the second person isn’t sure of the answer, he’s liable to go along with the first person’s guess. By then, even if the third person suspects another answer is right, she’s more liable to go along just because she assumes the first two together know more than she does. Thus begins an “informational cascade” as one person after another assumes that the rest can’t all be wrong...Because of this effect, groups are surprisingly prone to reach mistaken conclusions even when most of the people started out knowing better... If, say, 60 percent of a group’s members have been given information pointing them to the right answer (while the rest have information pointing to the wrong answer), there is still about a one-in-three chance that the group will cascade to a mistaken consensus.

Your Money and Your Brain

I thought I would pass on this brief commentary by Joe Nocera in the NY Times (PDF here), which emphasizes comments by Jason Zweig on how our emotional brains are not very well suited for making rational financial decisions.

A neural marker of consciousness

Here is the author's summary of an interesting article by Del Cul et al. in PlosBiology.

Understanding the neural mechanisms that distinguish between conscious and nonconscious processes is a crucial issue in cognitive neuroscience. In this study, we focused on the transition that causes a visual stimulus to cross the threshold to consciousness, i.e., visibility. We used a backward masking paradigm in which the visibility of a briefly presented stimulus (the “target”) is reduced by a second stimulus (the “mask”) presented shortly after this first stimulus. (Human participants report the visibility of the target.) When the delay between target and mask stimuli exceeds a threshold value, the masked stimulus becomes visible. Below this threshold, it remains nonvisible. During the task, we recorded electric brain activity from the scalp and reconstructed the cortical sources corresponding to this activity. Conscious perception of masked stimuli corresponded to activity in a broadly distributed fronto-parieto-temporal network, occurring from about 300 ms after stimulus presentation. We conclude that this late stage, which could be clearly separated from earlier neural events associated with subliminal processing and mask-target interactions, can be regarded as a marker of consciousness.


Figure: Top, depth of cortical processing: subliminal stimuli (left panel) should evoke a strong activation in extrastriate visual cortex, but their intensity should quickly decrease in higher visual areas; only conscious stimuli (right panel) should trigger a late surge of activation in a global prefronto-parietal network. Bottom, schematic time course of activation as a function of masking strength. Masking is expected to have little effect on early visual activation but to modulate the strength of activation in higher visual areas. Furthermore, there should be a nonlinear effect of masking strength in prefrontal cortex, with a similar late top-down activation peak occurring simultaneously in visual areas

Where the brain decides how much we are willing to pay.

Plassmann et al. show some brain correlates of our willingness to pay:

An essential component of every economic transaction is a willingness-to-pay (WTP) computation in which buyers calculate the maximum amount of financial resources that they are willing to give up in exchange for the object being sold. Despite its pervasiveness, little is known about how the brain makes this computation. We investigated the neural basis of the WTP computation by scanning hungry subjects' brains using functional magnetic resonance imaging while they placed real bids for the right to eat different foods. We found that activity in the medial orbitofrontal cortex and in the dorsolateral prefrontal cortex encodes subjects' WTP for the items. Our results support the hypothesis that the medial orbitofrontal cortex encodes the value of goals in decision making.
Neural correlates of WTP. A, B, Activity in the medial OFC and the DLPFC was positively correlated with WTP at the time of evaluation in the free trials more than in the forced trials.

What the F***?

An article by Steven Pinker on cursing.

Google's evil eye....

...thought I would link you to this interesting article.

Radiohead - the modern troubadours?

An article in the Arts section of today's NY Times temporarily stirred me out of my curmudgeonly old fart persona. Just as a federal jury in Minnesota last week decided that a mother found liable for copyright infringement for sharing music online should pay damages amounting to about $9,250 apiece for 24 songs, the British Rock group Radiohead has made their latest album "In Rainbows" available for download, letting their fans decide how much they want to pay for it. This has brought a new climax in the pricing pandemonium in the on-line music industry summed up by Radiohead manager Bryce Edge:

Digital technology has reintroduced the age of the troubadour. You are worth what people are prepared to give you in the digital age because they can get it for nothing.

In another departure from convention:

...the band declined to send out early copies of the music for reviewers and has not settled on a traditional single to push to radio stations. As a result, programmers are improvising. In San Francisco, for instance, the rock station KITS-FM, Live 105, has the entire album on its Web site (live105.com) and will let fans vote to determine which songs merit airplay.

I just checked out this site, and proceeded to buy and download the new album.

See-through frog...

Just the thing for a budding young scientist's holiday stocking...From the Oct. 3 issue of Nature Magazine:

Japanese biologists have created what they call the world's first see-through creature with four legs. (Some fish are naturally transparent.) The frogs started out as ordinary Japanese brown frogs (Rana japonica ), but crossing animals with recessive genes for light-coloured skin led to transparency. Internal organs, eggs and other normally concealed innards are all on full display in the new amphibian (pictured). Its creator, Masayuki Sumida of Hiroshima University, plans to seek a patent, according to the AFP news agency.

How we see over 10,000-fold changes in light intensity

The first 35 years of my professional life were spent studying the excitation and adaptation of photoreceptor cells. This motivates me to point out a beautiful piece of work by Dunn et al. showing how different circuits in the retina collaborate to let us see over an amazingly wide range of environmental light intensities. Here is their abstract, and figures from the paper showing the cells involved:

We see over an enormous range of mean light levels, greater than the range of output signals retinal neurons can produce. Even highlights and shadows within a single visual scene can differ approximately 10,000-fold in intensity—exceeding the range of distinct neural signals by a factor of approximately 100. The effectiveness of daylight vision under these conditions relies on at least two retinal mechanisms that adjust sensitivity in the approximately 200 ms intervals between saccades. One mechanism is in the cone photoreceptors (receptor adaptation)and the other is at a previously unknown location within the retinal circuitry that benefits from convergence of signals from multiple cones (post-receptor adaptation). Here we find that post-receptor adaptation occurs as signals are relayed from cone bipolar cells to ganglion cells. Furthermore, we find that the two adaptive mechanisms are essentially mutually exclusive: as light levels increase the main site of adaptation switches from the circuitry to the cones. These findings help explain how human cone vision encodes everyday scenes, and, more generally, how sensory systems handle the challenges posed by a diverse physical environment.

Figure 1: Midget and parasol ganglion cells adapt at lower backgrounds than L cones. Schematic of primate midget and parasol pathways with fluorescent images of cones in slice and ganglion cells in flat mount.

Figure 2: Post-receptor adaptation occurs in signal transfer from cone bipolar cells to ganglion cells. The figure shows fluorescent images of a midget cone bipolar cell (left panel) and a diffuse cone bipolar cell (right panel) in slice. OPL, outer plexiform layer; INL, inner nuclear layer; IPL, inner plexiform layer.

The wireless epidemic

Jon Kleinberg, best known for his work on web searching that paved the way for Google and the rest, writes on the threat posed by computer viruses and worms in our increasingly 'wireless' age, concluding that biological models of virus transmission are increasingly relevant for assessing the emerging threat. Here are a few clips:

Traditionally, computer viruses have propagated on networks that bear little resemblance to the networks of physical contact through which their biological counterparts spread. But a growing body of research shows that the increasing use of short-range wireless communication networks might cause the two models to converge....Diseases in plant populations, or animal diseases such as rabies, are heavily constrained by geographical proximity and the relatively fixed physical locations of the infected individuals. Models of these diseases have been extended using detailed data on patterns of travel within cities and by air worldwide in attempts to analyse disease outbreaks in human populations...Epidemics on the Internet are even more diverse. At the most general level, there is a distinction between computer viruses, which 'piggyback' on data exchanged between users, and computer worms, which more actively direct their own transmission through a network...

Very roughly.. one could view models of biological epidemics as rooted in spatial networks, and expanding into less spatial realms to model the technologies that have accelerated human travel. Meanwhile, research on cyber-epidemics has occupied the non-spatial end of the spectrum, with its diverse and far-flung connections, when modelling global communication technologies such as the Internet.

..the spread of short-range wireless communication technologies such as Bluetooth, and the emergence of worms that exploit these systems is disrupting this dichotomy by making possible computer-virus outbreaks whose progress closely tracks human mobility patterns. These types of wireless worm are designed to infect mobile devices such as cell phones, and then to continuously scan for other devices within a few tens of metres or less, looking for new targets. A computer virus thus becomes something you catch not necessarily from a compromised computer halfway around the world, but possibly from the person sitting next to you on a bus, or at a nearby table in a restaurant.

In assessing the risks of such attacks, and developing countermeasures against them, it is intriguing to contemplate how we might draw on expertise from the field of human epidemiology in understanding how contagion spreads...Analogies to biological epidemics can also be exploited for beneficial purposes, in the design of computer-network protocols. For mobile devices, epidemiology helps in dealing with the problem of intermittent connectivity: that the routing of traffic must conform to a dynamic and unpredictable network structure as the owners of mobile devices move around. The result is a growing interest in opportunistic routing, in which messages are passed between devices that come into physical proximity, with the goal of eventually reaching a specified recipient. The development of such protocols has drawn on detailed data concerning human mobility and contact patterns.

An acetylcholine receptor agonist improves cognition

The alpha-7 nicotinic acetylcholine receptor (nAChR) plays an important role in cognitive processes and may represent a drug target for treating cognitive deficits in neurodegenerative and psychiatric disorders. Bitner et al. study the effects of a particular AChR enhancer, or agonist, whose simple name is A-582941.

The figure shows the general structure of this class of molecules. A-582941 enhanced cognitive performance in behavioral assays including the monkey delayed matching-to-sample, rat social recognition, and mouse inhibitory avoidance models that capture domains of working memory, short-term recognition memory, and long-term memory consolidation, respectively. Their results demonstrate that alpha-7 nAChR agonism can lead to broad-spectrum efficacy in animal models at doses that enhance ERK1/2 (extracellular-signal regulated kinase) and CREB (cAMP response element-binding protein) phosphorylation/activation and may represent a mechanism that offers potential to improve cognitive deficits associated with neurodegenerative and psychiatric diseases, such as Alzheimer's disease and schizophrenia.