When the piece first premiered, critics When the piece first premiered, critics called it "repellent," "incomprehensible;" a "confusion." The audience didn't even call for an encore - a slight that threw the piece's composer into a rage. He sneered that the masses were ignorant "cattle" and "asses" - and by that point in his career, few men would've refused him the right to shout such things. Though Beethoven's hearing had been deteriorating for more than 20 years, many of his compositions still packed out auditoriums and commanded premium prices. Thunderclap symphonies like his Fifth and Ninth had carried the man's legend to the farthest courts of Europe.

So why, when his Great Fugue premiered in 1826, did audiences react with such horror and disgust? And why, more than a century after the fact, do critics now hail this same piece as a stroke of genius; an idea that arrived decades before its time? The most obvious reason, in both cases, is that the Fugue looks much more pleasing on paper than it sounds to an unprepared ear. Though it remains one of the most intricate and technically demanding pieces in all symphonic music, its wild dissonance has led even fans who praise its "perfection of detail" to caution that it's one of several Beethoven works that "should be excluded from performance."

In short, musical compositions that look beautifully balanced - even ornamental - on the page can produce painful cacophonies when translated into sound. Unlike printed text, which can only be read aloud one word at a time, musical notes have to sound beautiful - or meaningful, at least - not only when played one-at-a-time, but when played in rhythm with every other note in the same beat or measure.

This is also one of the most striking difference between hallucinations of music - which often take the form of familiar or catchy songs - and the more unusual cases of patients who hallucinate musical notation. As neurologist Oliver Sacks, author of the recent book Hallucinations, has discovered as he's studied these cases, even the most intricate and ornate hallucinated notation often translates to musical nonsense - which begs the questions of how and why people hallucinate such detailed notation in the first place.

Elaborate chaos

"Many hallucinations are accompanied by a feeling of astonishment," Sacks says. Although most of us don't feel like we have much direct control over even our everyday process of creativity, we can still usually tell the difference between imagining something - a musical composition or a narrative arc - and witnessing a performance of it. Hallucinations blur that boundary. A skeptical person might recognize that he or she is "seeing things," but those things still seem to have a life of their own - they weren't summoned out of the imagination by any conscious act of will; nor can they easily be willed away.

This creates an odd paradox for those who hallucinate: Many are well aware that the things they're seeing are "crazy" and irrational, but that doesn't make their visions any less insistent. Some fear - perhaps rightly - that their hallucinations will turn them into victims of the kind of mental health-based discrimination that remains all too common among coworkers, health care professionals and even family members, despite the fact that the rest of these patients' mental faculties remain fully intact. And so, when they learn of Dr. Sacks, they seek him out to share stories of the many unexpected places where their hallucinations appear.

One of Sacks' patients, for example, found that the embroidered border of his bathmat tended to transform into elaborate staves and clefs of music. He noticed a similar transformation in the text of his newspaper as he tried to read it, so one day he set his newspaper on his music stand and tried to play what he saw. The only problem, Sacks says, is that "the notation turned into different music every few seconds, and it kept altering." The patient, who also happens to be a professor of Sanskrit, discovered another odd twist when he studied his hallucinated notation more closely: The musical notes kept taking the shapes of letters in Sanskrit's traditional Devanagari alphabet.

Even patients who don't hallucinate notes in the shapes of ancient alphabets often find that the notation they hallucinate makes very little musical sense. "There's a superabundance of appoggiaturas and acciaccaturas and phrase marks and so forth," Sacks says, "which together compose a sort of nonsense." Thus, even if a patient finds that he or she can hold the hallucination in place long enough to study its details, those details may be just as chaotic as the swirling Sanskrit letter-notes. It's a perplexing thought: These ornate and intricate compositions springing up, fully formed, from some dark chamber of the brain - only to reveal themselves as what Sacks calls "completely unplayable and unsingable" chaos.

So if these hallucinations don't reflect some deep-seated process of unconscious composition, what does raise them from the depths? Do they hold any meaning at all?

Surreal autonomy

Ordinarily, when we hear a brilliant piece of music - or imagine it, or read it or play it - it's suffused with emotion and association, like a good book. But when it comes to hallucinations of musical notation, this often isn't the case. "On the whole, these hallucinations of musical notation - and, in fact, most lexical hallucinations - arouse no special feeling at all except surprise, and perhaps amusement or annoyance," Sacks says. Not only are these notation hallucinations musically nonsensical - they're also emotionally nonsensical. They don't recall any particular memory or feeling, but a jumble of loosely connected images and ideas.

Area IT (highlighted in orange here) plays a central role in integrating images of letters and musical notes with the sounds they represent.

In a strictly neurological sense, it's not too difficult to isolate the origins of hallucinations like these. PET and fMRI scans have helped us pin down brain areas whose activity is crucial for reading - and while reading music appears to involve many more brain regions than reading text, certain areas of the brain do seem to be central to both processes. One of these regions is the left inferotemporal cortex (also known as "area IT") which seems to help associate certain images, such as letters and notes, with the sounds they represent. People who suffer damage to this area often have trouble sight-reading text and music, though their emotional responses to the sounds they hear usually remain intact.

On the other hand, some brain areas known to be central to reading and understanding text - such as the left lingual and fusiform gyri - don't seem to respond to musical notation, handing over the interpretive reins to an area known as the left parietal-occipital junction when musical scores are involved. This functional shift, Sacks says, may hint that hallucinations of musical notation may spring from an entirely different brain pathway than other types of visual hallucinations - though we haven't yet confirmed or denied this with any fMRI studies of patients who hallucinate musical notation.

By way of analogy, Sacks points to Charles Bonnet syndrome - a brain disorder that produces hallucinations of seemingly random images from one's past (one Charles Bonnet sufferer, for example, relentlessly hallucinated the face of Kermit the Frog, a character who she said "meant nothing" to her). Though Sacks cautions that hallucinations of musical notation may not fall neatly into the Charles Bonnet category, they do share one significant feature in common with those images: Meaninglessness. It's as if a small group of brain regions decided to split off from the mainstream and throw a little surrealist party all on their own.

"In many of these hallucinations," Sacks explains, "instead of the usual chain of neural connections running from primary sensory cortex up to prefrontal cortex, we find that there's a little bit of activity in the middle areas of the brain - in inferotemporal or parietal cortex - that's taken on a life of its own, not informed by external reality, or necessarily by memory or emotion." In a sense, then, these hallucinations are entirely devoid of meaning - not only do they hold no emotional significance for the person who experiences them; they're often so incoherent that interpretation on any symbolic level is all but impossible.

But even so, one crucial question still remains: If these hallucinations are so meaningless, why does a brain go to the trouble of serving them up?

Anarchic significance

In some ways, hallucinations invite comparison with dreams. Both feature bizarre imagery that never seems to sit still; both arise unbidden from brain areas outside our conscious control; and both have been subjected to a wide range of explanations and interpretations over the years.

Before Freud, many scientists believed that dreams simply reflected semi-random neural firing that persisted in ancient brain areas like the pons and the medulla oblongata as we slept. Around the turn of the twentieth century, psychologists began insisting that dreams offered a sort of "royal road" to the repressed fears and desires of the unconscious mind - and this view persisted widely until the 1970s, when neurologists shored up the theory that dreams reflect our brains' replay and consolidation of short-term memories into long-term storage.

Although hallucinations differ from dreams in some fundamental ways - most notably, hallucinations occur during waking life, and they tend to produce astonishment rather than acceptance - they beg some of the same questions: Why do our brains choose to show us certain images as opposed to other ones? And can those images tell us anything about what's going on beyond the reach of our conscious minds?

"Those are very difficult questions to answer," Sacks says. On the one hand, he explains, "one can't avoid feeling that there may also be some anarchic or random elements" in hallucinations - but at the same time, "one would like to think that every hallucination must be determined by activity at some level in the brain." Whether that activity conveys meaningful insights about a person's mind is, of course, another question entirely - and "it's not always easy to derive a clear chain of causation from a symptom," Sacks says.

For example, anyone who hallucinates musical notation - or Sanskrit letters - must logically have seen the elements that compose the hallucination at some point in his or her life. Many patients who suffer from hallucinations of musical notation have spent decades of their lives reading, studying and playing music - and yet, strangely enough, this isn't always the case. Indeed, Sacks says, at least one of his patients who hallucinates musical notation can't read musical scores at all. "Simply having seen musical scores may be enough," he explains; "and almost everyone in our culture has seen music."

The significance those notes or letters hold in a person's memory, though, may not hold any direct bearing on the hallucinatory causes for which they're later recruited. Hallucinations of musical notation, like many Charles Bonnet-type hallucinations, tend to develop in people who are losing their eyesight. Thus, Sacks hypothesizes that these hallucinations - whether they take the form of arpeggios, lines of text, or imaginary frogs - may represent the attempts of a brain's failing visual system to fill in gaps in perception with whatever visual imagery happens to be available. '

"Normally," Sacks says, "in the act of perception or imagination, the entire visual system is engaged, activities at every functional level so seamlessly integrated as to defeat analysis." Disorders like blindness can provide clues about how this system breaks down - but only in "positive" disorders like hallucination, where the system does something "extra," do we gain insight into how our brains try to remedy these problems on their own.

The real meaning of hallucinations of musical notation, then, may not lie in the clefs, notes and ornamentations - or even, necessarily, in the fact that the imagery takes the form of musical notation at all. These hallucinations may not provide a "royal road" to a patient's unconscious mind - in fact, they may be entirely unrelated to a patient's level of musical ability or interest.

What they can tell us, though, is that our ability to perceive reality is a bit like a brilliant musical score: So intricate and multi-layered that its very complexity is often hidden from casual observation. And as Beethoven understood, it's in our moments of dissonance that the work's true intricacy begins to shine through.