Small world networks key to memory

By Philip Cohen If you recall this sentence a few seconds from now, you can thank a simple network of neurons for the experience. That is the conclusions of researchers who have built a computer model that can reproduce an important aspect of short-term memory. The key, they say, is that the neurons form a "small world" network. Small-world networks are surprisingly common. Human social networks, for example, famously connect any two people on Earth - or any actor to Kevin Bacon - in six steps or less. Properties like this have made them the focus of much research. It turns out that regardless of the size of these networks, any two points within them are always linked by only a small number of steps. Now it looks as if working memory, which allows short-term recall of fleetingly remembered information such as phone numbers, relies on the same property. This type of memory resides in an area at the front of the brain called the prefrontal cortex, which is involved in learning, planning and many higher cognitive functions.

Electrical bursts The late Patricia Goldman-Rakic of Yale University School of Medicine and others have suggested that neurons in this region might be able to switch between two stable states, a property called bistability. When storing a memory, neurons would participate in self-sustaining bursts of electrical activity. When not involved in memory storage, the neurons become quiet. Just how the brain controls this behaviour has puzzled neuroscientists. The prefrontal cortex is home to a wide variety of neurons with different properties, such as response to different chemical signals and the ability to activate or inhibit neighbouring neurons. So researchers have resorted to equally complex computer models. Now a team at Northwestern University in Evanston, Illinois, has reproduced the behaviour with the simplest of networks - by connecting it together to form a small world. "The philosophical conclusion is that connectivity matters," says team member Sara Solla. "Our model uses only a simple caricature of neurons, yet this network shows this working memory-like behaviour." Activating pulse Solla and colleagues Alex Roxin and Hermann Riecke began by creating a model of simple neurons that when activated would activate their neighbours for a brief period. This simple network behaved in a simple way: an activating pulse travelled through the network and then disappeared at the fringes. They then added short cuts to connect distant parts of the network: the key to a small world. They found that when 10 to 20 per cent of the neurons participated in short cuts, the network formed self-sustaining loops of activity. For example, region A would activate B through a short cut that would similarly trigger C. Finally, a short cut from C sparked A again, completing the loop. Intriguingly, a second strong activating pulse would shut the whole system down again. "So we get bistability," says Solla. The new work may be an important step towards a theory of how the brain works. "What I really like about this system is how simple it is," says John White, a biomedical engineer at Boston University. "That means it might be possible to move from computer models to a rigorous mathematical description, which would really advance the field." Using six degrees of separation (November 2000) The notion of 'six degrees of separation' is that everybody in the world is linked to everybody else by no more than six steps of acquaintance. But, while this makes an interesting topic for filler articles in a newspaper, it has been of very little real use, until now, when a Cornell sociologist has shown how the concept can be used as a scientific sampling method for finding and studying 'hidden populations,' from drug users to jazz musicians. According to an article by Douglas Heckathorn and Joan Jeffri in the November issue of Poetics, the method they have outlined can be used to obtain a scientifically valid, representative sample of populations that cannot be identified using traditional sampling methods. These would include, for example, drug addicts, HIV-infected individuals, the homeless, runaway youths, gays and lesbians, poets and other creative workers who would be hard to spot in a crowd. In simple terms, there are no lists available or associations of runaway youths, but individuals in a group know each other, and the sampling method takes advantage that. According to Heckathorn, ''As we gather information during the sampling process of referrals, we look at the degree to which people tend to recruit those similar to them. Then, we can mathematically correct for the non-randomness and project what the sample would have been had there been no biases.'' The survey method was developed to study a peer intervention program with drug users in Connecticut, Chicago and Russia. Now Heckathorn is to apply it for a study of jazz musicians for the (US) National Endowment for the Arts (NEA). The study seeks to determine the socio-economic profiles of these musicians, such as whether they have health and life insurance, data on copyright protection, use and abuse of new technologies, their level of income from jazz and jazz-related activities, number of jobs the musicians need to survive, their experiences with mentors, teaching, distribution, marketing, and management and retirement. ]azz artists, he says, '' . . . exist in a kind of no-man's land, where earning a living from jazz is almost impossible and where even individual support like the Jazz Masters awards from the NEA are not enough to offset the hand-to-mouth existence of most jazz musicians.'' So the aim is to obtain a statistically valid sample, and then to seek to determine the musicians' current situation and most pressing needs. For information on Douglas Heckathorn, see http://www.soc.cornell.edu/faculty/heckathorn.shtml, and for more information on the sampling method, see http://www.people.cornell.edu/pages/ddh22/pubs.html ©WebsterWorld Pty Ltd/contributors 2002