Usually the best source of information is going to be the logs on the target of the attack. A device like a router or a Linux/Unix server is going to show who last logged on and from what IP. Unfortunately, if the target was itself Windows-based, the quality of the logged information deteriorates markedly.
Since you believe it was a Windows-based workstation, it is likely that the IP address would not be spoofed and could lead you back to the workstation from which the hack was launched.You might also be able to infer something from the NT Security logs (unusual logon time, perhaps) and go from there.
But as far as I know, use of telnet is not an auditable event and using telnet from the command prompt does not leave much of a footprint on the machine.
Without that information, you're looking for the proverbial needle in the haystack and unless you are logging all telnet traffic across a router or have an IDS system up and running, it's going to be a tough case to crack and will get tougher for each day that goes by.
paul