The U.S. stock market plunged Monday, with the Dow Jones falling nearly 1,600 points at one point, the biggest single day drop in its history. It then turned around and regained 567 points on Tuesday. What the remainder of the week holds in store is anyone’s guess.

Many experts had been forecasting a decline for months after a prolonged upswing resulted in a series of record highs. Several factors are likely to have been involved. The Bureau of Labor Statistics January jobs report, released on Friday, was almost certainly one of them. It generated worries about inflation and bond yields, together with concern The Federal Reserve may raise interest rates faster than expected—events that may have “spooked” the markets.

Markets translate the decisions of millions of people into a price for a stock or bond. Like a spooked crowd in a public place, investors tend at times to run in the same direction—let’s all play the lottery or let’s escape the burning movie theater.

The work of visionaries such as Nobel laureates Richard Thaler and Daniel Kahneman has demonstrated humans do not operate as rational agents, as assumed by classical economics. From this realization have emerged disciplines such as behavioral economics, neuroeconomics and the like.

After the downs and ups of the markets in recent days, journalist Simon Makin spoke to neuroeconomist and psychiatrist Richard Peterson, managing director of MarketPsych research, whose most recent book, Trading on Sentiment: The Power of Minds over Markets, explores the role of crowd psychology in securities pricing, using an analysis of social media data about what underlies violent downs and ups of trading.

[An edited transcript of the interview follows.]

What does yesterday’s market downturn say about the fickleness of market psychology and how people react when they perceive a threat to their wealth?
After a long period of no dislocations in the stock market, there was a rise in risk perceptions (or fear) over the past few weeks. While prices were rising it wasn't widely noticed, but once they started to decline it triggered a rapid fear response.

The large January declines in bonds and speculative assets (like cryptocurrencies) likely triggered activity in [investors’] anterior insulae (part of the brain’s threat-detection system). Once investors are emotionally primed by these losses, their emotional frame changes to one of scanning for further negative cues. Our brains begin to search the news for more negative information. When we find it, we sense threat and our amydgalas are activated—comparing the current threat to memories of past threatening experiences. Shorter-term traders then take action to preserve their highly leveraged gains.

We [MarketPsych] have evidence for this increase in fear before the sell-off from data we collected on media sentiment, using all references to fear—“worries,” “concerns,” “caution,” etcetera—about stocks in the Nasdaq 100 gathered from 800 global financial social media sites. The short-term average of this measure had been rising rapidly for about two weeks. This means investors were emotionally primed with fear, an emotion known to correlate with lowering bid and ask prices in experimental markets.

Social media wasn't such a presence in, say, 2008. Do you think that may be having an impact on the volatility of markets?
Yes, we see evidence of this in our data. Bubble-ometer measures [one of the barometers developed by MarketPysch to estimate market vulnerability] use net positive emotionality and expectations gathered from media output to indicate increasing bubble risk. This shows the level of speculative activity spiking higher—indicating bubbly and speculative language—then rolling over before the sell-off. This is more pronounced using material from social media only, compared to news only, and tracks the timing more accurately. Investors who survive on social media tend to be fairly “wise” and more accurate than news media, according to our research.

Does a day like Monday in the markets demonstrate anything about the rationality, or otherwise, of trading securities?
Yes, it shows that market movements can be irrational—spiking higher in the prior months and now crashing down abruptly. In finance research there are patterns called “overreaction” and “underreaction” to information. During rising trends, investors are said to be slowly reacting (underreacting) to the accumulation of good news. When there are sharp sell-offs and the market then later bounces higher, investors are said to have overreacted to negative news.

Cognitively, it appears investors slowly process boring, good information, leading to underreaction, but overreact to dramatic, vivid information (like sharp sell-offs) setting the market up for a rebound.

The global economy is very healthy, although the markets are thought to be at fairly high levels. What is it about crowd psychology that makes people overreact?
It’s herd behavior. When we are uncertain about why prices are moving, we look to others’ behavior to give us direction. In this case, we look to price movement as a proxy for others who know better, as an indicator of the knowledge they possess that we don’t. As prices decline, this creates a positive feedback loop—as we all sell, thinking others know better—which drives prices lower and lower.

Would you say automated trading algorithms were involved?
Yes, but I don’t know to what extent. There are trend-following and volume-weighted execution algorithms that create the exponential spikes and dips in prices. Beyond that, it’s difficult to get details due to the secrecy of the industry.

Do you think this is the end of the “Trump Bump”?
Long-term, no, unless some new issue arises such as higher bond prices. There’s a bigger financial issue for Trump inflation and higher bond prices. He’s a credit guy and likes to spend. He also has a history of spending himself into bankruptcy.

Does this have the same hallmarks as other market panics?
No, it’s simply a correction in the midst of a massive bull market. In fact, it’s healthy for the longer-term bull market as it “culls the herd.” There’s enormous wealth and liquidity outside of stocks waiting to get back in that could propel this market higher. We will hopefully be back into an era of higher volatility, which is generally healthier.