The biographical description that follows is adapted from the 1988 APA Distinguished Scientific Contribution Award (published as: Anon. Biography. American Psychologist, 1989, 44, 626-628). The biography has been updated with the description of Professor Sperlingís more recent and ongoing research. Sperling is a Fellow of the National Academy of Sciences and the American Academy of Arts and Sciences and the Society for Experimental Psychologists, and a recipient of the APA Distinguished Scientific Contribution Award (1988), the Howard Crosby Warren Medal (1996) of the Society for Experimental Psychologists, The Edgar D. Tillyer Award (2002) of the Optical Society of America, and the Helmholtz Award (2004) of the International Neural Network Society.

George Sperling was educated in the public schools of New York City. He went to the University of Michigan to study science, majoring simultaneously in biology, chemistry, physics, and mathematics, receiving a BS in l955. After an MA in psychology from Columbia University in 1956, he went on to Harvard. During the summer of 1957, George Miller arranged for Sperling to borrow a tachistoscope in order to carry out an experiment he had proposed in a student report. Sperling used a method of partial report to measure the time course of visual persistence (sensory memory). Subsequently, Neisser (1966) renamed it "iconic memory." With S. S. Stevensí help, the summer experiments became Sperling's doctoral thesis in 1959.

The summer of 1958 was spent at Bell Laboratories. The previous summer's experiments were repeated, and Sperling first used post-stimulus visual masking to measure the rate of information extraction from visual materials, uncontaminated by visual persistence. (His post-stimulus mask was a stimulus consisting of visual noise that overwrote and thereby terminated the persistence.) Sperling mailed photographic copies of his masking stimulus to numerous colleagues but enticed only his colleagues across the hall, Averbach and Coriell. The method was published in 1963, and now, of course, post-stimulus masking by visual noise is universally used.

Another observation relative to his thesis was that subjects occasionally made errors indicating acoustic confusions. Sperling (1963, 1967) theorized that items were transferred from very short-term visual memory to an auditory memory. With Roseanne Speelman, he demonstrated the importance of acoustic confusions in short-term memory in recall of sequences such as BPDTCV and developed a phonemic model of short-term auditory memory that quantitatively predicted the memory impairment for acoustically confusable items (1968).

Another early Bell Labs experiment required Sperling to program a primitive Honeywell computer to generate rapid sequences of character arrays (now called RSVP, rapid serial visual presentation). These rapid sequences again simulated the sequence of images produced on the retinal by natural eye movements during visual search. Searching for a numeral target among letter distracters was twice as fast in simulated displays as in natural search. The highest search rates were achieved when arrays occurred at 20-25 per second (8x faster than in natural search); 4-5 locations were searched concurrently.

Sperling worked at Bell Laboratories full-time at first and then jointly with an appointment at New York University for 26 years beginning in 1959. In the early 1960s, Sperling proposed the auditory synchronization method for measuring visual persistence duration. This method requires the subject to judge the synchrony of a click and the onset or termination of a light flash. Many years later, with Erich Weichselgartner, the method was extended to measure not just the moment at which visual persistence ceased but also the entire rise and fall of the temporal brightness function.

What had attracted Sperling to psychology originally was the prospect that the quantitative methods and formal theories used by physicists to probe and describe the nature of matter could be adapted to yield equally powerful measurements and theories of mental micro-processes. Although the specific content of this early work was important--for example, in initiating the study of what is now called sensory memory--the style of more atomic theoretical conceptualizations had a much greater influence. Some measure of the impact of Sperling's 1960s work can be garnered from citations to the three articles: the Psychological Monograph that introduced partial report to measure "iconic memory" in 1960, his 1963 article that introduced masking by post-stimulus visual noise to measure the rate of information acquisition, and his 1967 article on "Successive approximations to a model for short-term visual memory that introduced the auditory synchronization method. All were citation classics and were among the 99 most cited articles in psychology according to a 1970s survey.

Sperling's first publication in visual psychophysics was "Negative Afterimages Without Prior Positive Images," an explanation of Bidwell's top. Subsequently, he published mathematical models for adaptation and flicker (in 1968 with M. M. Sondhi), contrast detection, binocular vision, and motion perception. The 1968 paper with Sondhi introduced and analyzed feed-forward shunting inhibition, a neural process that is now regarded as one of the most universal in neural processing and is a component of many subsequent neural models.

Following a query by John Pierce, a distinguished Bell Labs physicist, Sperling investigated the spatial localization of flashes that occurred during brief saccadic eye movements. He also built an elaborate apparatus to produce the same continuous display on a stationary retina as an eye movement produced on the moving retina. For eye movements of 4 deg, the spatial mislocalizations of the flash, which began before eye movements and continued afterwards, were quite similar in both the real and the simulated eye movements. For larger deg eye movements, localization was better in real versus simulated displays. The data were nicely captured by a theory that proposed the perceived eye movement was slightly slower and began slightly before the physical eye movement. In simulations of eye movements that were 5x slower than naturally occurring eye movements, the perceived eye movements still began earlier but were now faster than the simulated movements, thereby producing a phenomenon now called the "flash-lag" effect. Abstracts of these findings were published in 1965 and 1966, the 1960s data were first published in 1990.

Sperling's 1970 binocular model was cast in the form of a physical and a neural theory. The physical theory predates Thom's catastrophe theory and probably was the first application of potential theory in psychology to describe the phenomena of path dependence and multiple stable states. The neural model embodies several important innovations (a winner-take-all network, cooperation-competition interaction, and different levels of spatial resolution) that were adopted by subsequent modelers (Julesz, Grossberg, Nelson, Dev, Marr & Poggio, and others). Returning to binocular vision, in 2006 Ding and Sperling formulated a physiologically based model that precisely described the proportions in which disparate images from the two eyes were combined in binocular vision. Returning to binocular vision in 2006, Ding and Sperling formulated a physiologically based model that precisely described the mechanisms that determined how disparate images from the two eyes were combined in binocular vision.

Van Santen and Sperling (1984, 1985) proposed a mathematical theory of human motion detection. Their detectors were elaborations of Reichardt detectors that extracted Fourier components in dynamic motion stimuli. Sperling's empirical insight was that to study motion systems required passing visual stimuli through the early stages of vision without distortion, and this required stimuli with maximum contrasts of less than 5%. The second of these papers shows that two subsequently proposed models of visual motion perception, Adelsonís and Bergen's motion energy model and Watsonís and Ahumada's Hilbert space model are computationally equivalent to the Reichardt model.

Chubb and Sperling (1988) discovered a whole gamut of moving stimuli which they called "driftbalanced" that were invisible to Reichardt detectors and which required rectification to expose their latent motion. Ultimately, they showed that both motion and texture perception were served by two parallel computational systems: a Fourier and a rectifying system, that are now known as first- and second-order systems. Some stimuli similar to some of their second-order stimuli had been observed before. What Chubb and Sperling added was a formal theory that showed how such stimuli were processed and how to generate new ones.

In 1995, Lu & Sperling elaborated van Santen and Sperling's 1984 "motion pedestal paradigm" and introduced a new phase cancellation paradigm to demonstrate a third-order motion system that was distinct from first- and second-order motion. Third-order motion depended on salience, i.e., on what makes figure distinct from ground. Salience was determined by both bottom-up stimulus driven processes and top-down attentional processes. They demonstrated (third-order) motion stimuli in which the direction of apparent motion depended on the feature (e.g., black or white) to which the observer attended (Nature, 1995). Subsequently, with Blaser and Lu, Sperling turned the paradigm around and used a third-order motion paradigm to measure the strength of attention. The computational model for these experiments introduced a critical distinction between the effect of attention on salience and the effect of attention on appearance and subsequent computations. With two students, Tseng and Gobell, Sperling used a third-order motion paradigm to delicately measure the attentional sensitization produced by attending to a color and to show, surprisingly, that this sensitization could persist unabated for a month (Nature, 2004).

Starting quite early in Sperling's career, his research on sensory memory evolved naturally to studies of visual search and attention. Using a computer to present stimuli for search at a rate of 25 displays per second, the eye-movement controlled search rate could be doubled (Sperling, Budiansky, Spivak, & Johnson, 1971). With M. J. Melchner, Sperling (1978) published the first empirical Attention Operating Characteristic. In 1984, Sperling demonstrated the computational equivalence of signal detection theory (e.g., the receiver operating characteristic) and attentional resource theory (the attention operating characteristic).

Sperling and A. Reeves (1980) developed a new paradigm that enabled them to describe the entire reaction-time distribution of an unobservable shift of visual attention. Weichselgartner and Sperling (1987) measured automatic and controlled attention shifts initiated consecutively by the same stimulus. Reeves and Sperling (1986) developed a computational theory of attentional gating to account for their earlier results. The attention gating theory was elaborated by Sperling and Weichselgartner in an "episodic theory" of the dynamics of attention. Attention episodes were shown to be discrete, not continuous, with long dwell times and quick shifts independent of the spatial distance shifted. Shih and Sperling, in a Herculean effort, developed a new paradigm to measure spatio-temporal shifts of visual attention. They show that attention windows opened after about 0.15 sec in response to a cue, remained open for about 0.3 sec, and that information was gathered with the same time course from all points within the window. In Sperling's 1960 monograph, there was no computational theory of attention or iconic memory. Sperling and Reeves (1986) predict about 200 data points per subject from their simple attention-gating theory. Sperling & Shih (2002) collected tens of thousands of points and predicted over 400 data points from three complex experiments for each subject. Their theory encompasses the major paradigms of attention. Hsu, Scofield, and Sperling developed and tested a model of spatial attention that predicts the actually achieved distribution of spatial attention in response to any (they claim) of 1.5 10^42 possible requested attention distributions.

Deaf people who communicate by a manual sign language (such as American Sign Language, ASL) cannot use the telephone network to communicate dynamic images because its capacity is too small for ordinary television transmission. In 1980 Sperling made the first measurements of the minimum bandwidth needed to perceive ASL. Several laboratories jumped into the race to be first to develop image-processed ASL that could be sent over telephone lines. In 1985, Sperling, Landy, Cohen, and Pavel tested all of the then-available compression methods and demonstrated several feasible analog and digital methods. Although their research had little impact in the USA, England and Sweden developed and now employ systems based on these principles.

Sperling has had an enormous influence on contemporary cognitive science by consistently showing the way to develop careful, analytic experiments that enable, formal computational theories to describe the brain algorithms that underlie perception and cognition. His research provides a proof by example of the way that science in our field should and can be done.